Skip to content

Releases

v1.6.6

Choose a tag to compare

@github-actions github-actions released this 17 Nov 10:23

Ready-to-run builds of the ChromIQ X-Rite instrument information
collector: for the i1iO table, since 1.5 for handheld instruments
(i1Pro, i1Pro 2, i1Pro 3 / 3 Plus, ColorMunki, i1Studio), and since
1.6 for the i1iSis chart readers (i1iSis, i1iSis XL, i1iSis 2,
i1iSis 2 XL).

Just want to run it? Download the file for your system:

System File
Windows chromiq-io-probe-windows.exe
macOS (Intel and Apple Silicon) chromiq-io-probe-macos
Linux chromiq-io-probe-linux

No Python needed. It changes nothing on the instrument or the
computer
: it never sends anything to the instrument or the table,
cannot move the i1iO's arm, and changes no settings. The optional
recording only listens (on Windows through the free USBPcap, which
you install yourself if you want to record).

What's new in 1.6.5

  • Windows: "USB recorder attached to your instrument: yes /
    no / could not tell"
    is checked before the first step. If
    the recorder (USBPcap) is not attached, the tool says what to
    do (restart Windows with "Restart", then plug the instrument
    in) before you spend time on the steps.
  • After the first step, if only the instrument's USB
    description was recorded, the tool says so at once and offers
    to stop, instead of after all nine steps.
  • The report says what was seen.

What's new in 1.6.4

About older reports: reports made with a version before
1.6.3 contain codes that can be traced back to short serial
numbers. Please do not post such a report publicly, and delete
one you already posted if you can. --clean cannot change those
old codes.

  • Linux: a USB stick plugged into the instrument's old port
    later no longer shows its name in the report: a port counts
    as the instrument's only until it is unplugged.
  • Known instruments are named (for example a DTP20 or an
    i1iO 3 table) instead of "unknown X-Rite device".
  • macOS: the few details kept for other USB devices (ids,
    version, speed) are kept only when they look like numbers of
    that kind; anything else is not shown.

What's new in 1.6.3

  • When a step keeps nothing: a tester's screenshot showed a
    Windows session in which every step stayed empty; the cause
    is not known
    . The tool still keeps nothing it cannot
    attribute. It now says what such a recording held (counts,
    kinds of device and USB ids only, never contents) and what to
    try, one action per line (replug the instrument; if still
    empty, restart Windows with "Restart", not "Shut down"). A
    session in which no step kept anything ends with one small
    SEND-THIS-<date>-report-only.zip: only the report and
    session.json, no recording.
  • A handheld session keeps only its own instrument family
    (the i1Pro and the i1Pro 2 count as one), as the i1iSis
    session does; an unknown X-Rite device only with a note.
  • The wrong instrument chosen: when a step holds only another
    instrument's traffic, the tool says so first, names the number
    to choose, and offers to stop (no replug advice then).
  • The report's device lists show other USB devices without
    their names
    (ids and kind only; an allow-list); --clean
    does the same for older reports; --no-redact shows
    everything and says so.
  • The short codes that replace serial numbers are salted per
    run
    and can no longer be traced back by trying numbers.
  • Steps shown after another device's traffic was removed print
    no decoded text; no decoded string carries control
    characters; 0xCF is shown as "clock mode N".
  • Analysis: ArgyllCMS does send the i1Pro 3's 0x3B (it was
    listed as never sent); the i1Pro 3's ruler travel per scan;
    every swipe by its length; the dual scan is not compared in
    "only with the ruler".
  • Steps: after the single patch, switch back to strip scanning;
    a lights answer "no" is asked once more.

What's new in 1.6.2

A privacy fix. Please read this if you made a recording with
1.6.1 or older.

  • Versions up to 1.6.1 could keep traffic of another USB
    device
    that shared the instrument's USB hub (for example a
    USB disk) in the files meant to hold only the instrument's
    traffic, and in analysis.txt and the report.
  • 1.6.2 leaves out traffic at the instrument's address on
    endpoints the instrument does not have (the case found), other
    devices' kinds of traffic (disk, keyboard, mouse, hub), and
    monitor calibrators. When another device was also seen on one
    of the instrument's own endpoints, the analysis shows no
    contents of that step and asks you to record it again.
  • It cannot tell apart a device that was given the
    instrument's address and uses the same endpoint numbers (for
    example a USB 3 "UAS" disk, a printer, a phone or a USB-serial
    adapter), nor another device's vendor requests.
  • What to do: update; clean old files with
    --clean SEND-THIS-....zip (or delete them); and while you
    record, unplug every other USB device you can, or at least
    those on the same hub or dock. If you already sent us an older
    file, tell us and we will delete it wherever we can, and treat
    it as private meanwhile.
  • Clearer steps: one action per line; wait for the ready signal
    before moving; "slow" means about twice as long as the normal
    scan, counted. The measurement condition is asked.
  • Analysis: lights in words, the ruler's motion per scan, fast
    and slow judged by the ruler.

What's new in 1.6.1

  • --analyze with several recordings of different
    instruments
    (for example an i1Pro 2 recording and an i1iSis
    recording) no longer reads them as one. It lists which
    instrument each file shows and asks you to analyse each
    instrument's recordings on their own. This closes the known
    limit of 1.6.0. Session folders, and what every session keeps
    in the copy to send, are exactly as in 1.6.0.
  • The i1iSis part of the README now says right there that the
    session has not been tried with a real i1iSis yet.

What's new in 1.6.0

The i1iSis chart readers. If you own an i1iSis, i1iSis XL,
i1iSis 2 or i1iSis 2 XL, the tool now offers a guided recording
session for it (or start it with --isis). It has not been
tried with a real i1iSis yet
: everything was tested with
invented recordings only, so your first session is the real
test. Nothing is promised: ChromIQ cannot use an i1iSis, and
whether it ever can is a separate question.

  • Six short steps in i1Profiler: start it and leave it idle;
    get the i1iSis ready for a page without feeding it; feed one
    page in single scan; the same page in dual scan; the same page
    a little crooked; a plain sheet that fails on purpose. After
    each step one extra question: how did the page move?
  • Switch the i1iSis on before plugging in its USB cable
    (X-Rite's own note). On Windows 11 25H2 use i1Profiler 3.8.8
    or newer, which X-Rite says fixed the i1iSis drivers there.
  • Any printed i1iSis chart page works (its colours do not
    matter for the recording); the tool says how to make one in
    i1Profiler and ships no chart.
  • Only the i1iSis is kept. When a recording shows the
    i1iSis, the copy to send holds the i1iSis and nothing else
    (no keyboard, no monitor calibrator). Without an i1iSis in it,
    only an X-Rite device the tool does not know is kept, and the
    tool says so. Without a USB description nothing is kept.
  • The analysis puts the steps side by side, says whether the
    i1iSis seems to talk through a USB-to-serial chip like the
    i1iO, and shows the first lines the program sent. The
    i1iSis's answers are never shown, only their sizes.
  • The read-only promise, tested as far as a test can: a test
    runs a whole simulated session while it watches and blocks the
    usual ways of opening files and starting programs (open,
    io.open, os.open, pathlib, subprocess, os.system,
    os.popen, os.posix_spawn, os.spawn*, os.exec*,
    os.fork, os.forkpty, the pty module, multiprocessing,
    and ctypes, which the tool may use only through Windows'
    kernel32 and shell32). It fails if the tool opens a device,
    writes outside its folders, or starts a program other than the
    ones it names with the arguments it really builds
    (PowerShell: the exact scripts); eighteen deliberately broken
    copies are all caught. A test cannot prove that nothing could
    slip through some other way.
  • Known limit: two plain files given to --analyze
    together that hold devices at the same USB address can be
    named after the wrong instrument family, and what the program
    sent to the other instrument can then appear among the lines
    sent to the i1iSis (the devices' answers stay hidden; the
    copies to send are not affected). Fixed in 1.6.1.
  • An independent review before the release found, and this
    release fixes: a device guessed from i1Pro commands could end
    up in the i1iSis copy when the recording had no USB
    description; the analysis of a raw recording could list a
    monitor calibrator as the instrument; a restart as
    administrator on Windows forgot --isis and --steps; some
    texts still spoke to i1Pro owners in an i1iSis session.
    A second review found that the first fixes made an i1iO table
    recording keep only an i1iSis on the same bus: now every
    session except the i1iSis keeps exactly what 1.5.5 kept,
    checked against the tagged 1.5.5 itself. A third review found
    that the analysis of a raw i1iSis recording with a monitor
    calibrator beside it mixed both; fixed, and the build's smoke
    steps are now also run locally before every release.

What's new in 1.5.5

Fixes from the first real Windows recording. Version 1.5.4
made a recording on Windows possible for the first time: a tester
recorded seven of nine steps with an i1Pro 3. Thank you! What his
session showed, and what 1.5.5 changes:

  • The Enter key: keys pressed while the tool is busy are now
    thrown away before every question (in 1.5.4 they answered the
    next questions, so two steps starte...
Read more

v1.6.5

Choose a tag to compare

@github-actions github-actions released this 14 Nov 14:43

Ready-to-run builds of the ChromIQ X-Rite instrument information
collector: for the i1iO table, since 1.5 for handheld instruments
(i1Pro, i1Pro 2, i1Pro 3 / 3 Plus, ColorMunki, i1Studio), and since
1.6 for the i1iSis chart readers (i1iSis, i1iSis XL, i1iSis 2,
i1iSis 2 XL).

Just want to run it? Download the file for your system:

System File
Windows chromiq-io-probe-windows.exe
macOS (Intel and Apple Silicon) chromiq-io-probe-macos
Linux chromiq-io-probe-linux

No Python needed. It changes nothing on the instrument or the
computer
: it never sends anything to the instrument or the table,
cannot move the i1iO's arm, and changes no settings. The optional
recording only listens (on Windows through the free USBPcap, which
you install yourself if you want to record).

What's new in 1.6.5

From a tester's Windows session: nine steps over 42 minutes kept nothing,
because USBPcap, the recorder, was not attached to the instrument. Most
likely Windows had set the instrument up again without it being plugged in
afresh, for example after its driver was changed. The tool now notices this
early.

  • Before the first step (Windows): the check of the PC now says
    "USB recorder attached to your instrument: yes", "no" or "could not
    tell". If it is "no", the tool explains it and lists what to do: close
    the tool, unplug the instrument, restart Windows with "Restart" (not
    "Shut down"), plug the instrument in afterwards, and start the tool
    again. Pressing Enter stops there; typing "go on" records anyway. "Could
    not tell" never stops anything. The i1iO table session checks the table
    and the instrument on it. In an "another X-Rite instrument" session one
    checked device the recorder sees is enough; any other device it does not
    see is named, without stopping.
  • After the first step: if it recorded only the instrument's USB
    description and nothing the instrument did, the tool says so right away,
    instead of after all nine steps. Two things look exactly alike here, and
    the tool says both: the measuring program did not talk to the
    instrument, or the recorder did not see it. It tells you what to check
    and offers to stop (Enter), to record the step again (r), or to go on.
  • The report (and the small report-only zip) says what was seen: the
    recorder line and, if it happened, that the first step held only the
    USB description.

What's new in 1.6.4

About older reports: reports made with a version before
1.6.3 contain codes that can be traced back to short serial
numbers. Please do not post such a report publicly, and delete
one you already posted if you can. --clean cannot change those
old codes.

  • Linux: a USB stick plugged into the instrument's old port
    later no longer shows its name in the report: a port counts
    as the instrument's only until it is unplugged. A kernel line
    that names no USB port is kept only when it carries the
    instrument's USB id, not because of a name in it.
  • Known instruments are named (for example a DTP20 or an
    i1iO 3 table) instead of "unknown X-Rite device". What the
    tool advises you to do is the same as before, except for an
    i1iO 3 table: like the other i1iO tables it now says to
    choose 1 (record an i1iO table) at the first question.
  • macOS: the few details kept for other USB devices (ids,
    version, speed) are kept only when they look like numbers of
    that kind; anything else is not shown.

What's new in 1.6.3

  • When a step keeps nothing: a tester's screenshot showed a
    Windows session in which every step stayed empty; the cause
    is not known
    . The tool still keeps nothing it cannot
    attribute. It now says what such a recording held (counts,
    kinds of device and USB ids only, never contents) and what to
    try, one action per line (replug the instrument; if still
    empty, restart Windows with "Restart", not "Shut down"). A
    session in which no step kept anything ends with one small
    SEND-THIS-<date>-report-only.zip: only the report and
    session.json, no recording.
  • A handheld session keeps only its own instrument family
    (the i1Pro and the i1Pro 2 count as one), as the i1iSis
    session does; an unknown X-Rite device only with a note.
  • The wrong instrument chosen: when a step holds only another
    instrument's traffic, the tool says so first, names the number
    to choose, and offers to stop (no replug advice then).
  • The report's device lists show other USB devices without
    their names
    (ids and kind only; an allow-list); --clean
    does the same for older reports; --no-redact shows
    everything and says so.
  • The short codes that replace serial numbers are salted per
    run
    and can no longer be traced back by trying numbers.
  • Steps shown after another device's traffic was removed print
    no decoded text; no decoded string carries control
    characters; 0xCF is shown as "clock mode N".
  • Analysis: ArgyllCMS does send the i1Pro 3's 0x3B (it was
    listed as never sent); the i1Pro 3's ruler travel per scan;
    every swipe by its length; the dual scan is not compared in
    "only with the ruler".
  • Steps: after the single patch, switch back to strip scanning;
    a lights answer "no" is asked once more.

What's new in 1.6.2

A privacy fix. Please read this if you made a recording with
1.6.1 or older.

  • Versions up to 1.6.1 could keep traffic of another USB
    device
    that shared the instrument's USB hub (for example a
    USB disk) in the files meant to hold only the instrument's
    traffic, and in analysis.txt and the report.
  • 1.6.2 leaves out traffic at the instrument's address on
    endpoints the instrument does not have (the case found), other
    devices' kinds of traffic (disk, keyboard, mouse, hub), and
    monitor calibrators. When another device was also seen on one
    of the instrument's own endpoints, the analysis shows no
    contents of that step and asks you to record it again.
  • It cannot tell apart a device that was given the
    instrument's address and uses the same endpoint numbers (for
    example a USB 3 "UAS" disk, a printer, a phone or a USB-serial
    adapter), nor another device's vendor requests.
  • What to do: update; clean old files with
    --clean SEND-THIS-....zip (or delete them); and while you
    record, unplug every other USB device you can, or at least
    those on the same hub or dock. If you already sent us an older
    file, tell us and we will delete it wherever we can, and treat
    it as private meanwhile.
  • Clearer steps: one action per line; wait for the ready signal
    before moving; "slow" means about twice as long as the normal
    scan, counted. The measurement condition is asked.
  • Analysis: lights in words, the ruler's motion per scan, fast
    and slow judged by the ruler.

What's new in 1.6.1

  • --analyze with several recordings of different
    instruments
    (for example an i1Pro 2 recording and an i1iSis
    recording) no longer reads them as one. It lists which
    instrument each file shows and asks you to analyse each
    instrument's recordings on their own. This closes the known
    limit of 1.6.0. Session folders, and what every session keeps
    in the copy to send, are exactly as in 1.6.0.
  • The i1iSis part of the README now says right there that the
    session has not been tried with a real i1iSis yet.

What's new in 1.6.0

The i1iSis chart readers. If you own an i1iSis, i1iSis XL,
i1iSis 2 or i1iSis 2 XL, the tool now offers a guided recording
session for it (or start it with --isis). It has not been
tried with a real i1iSis yet
: everything was tested with
invented recordings only, so your first session is the real
test. Nothing is promised: ChromIQ cannot use an i1iSis, and
whether it ever can is a separate question.

  • Six short steps in i1Profiler: start it and leave it idle;
    get the i1iSis ready for a page without feeding it; feed one
    page in single scan; the same page in dual scan; the same page
    a little crooked; a plain sheet that fails on purpose. After
    each step one extra question: how did the page move?
  • Switch the i1iSis on before plugging in its USB cable
    (X-Rite's own note). On Windows 11 25H2 use i1Profiler 3.8.8
    or newer, which X-Rite says fixed the i1iSis drivers there.
  • Any printed i1iSis chart page works (its colours do not
    matter for the recording); the tool says how to make one in
    i1Profiler and ships no chart.
  • Only the i1iSis is kept. When a recording shows the
    i1iSis, the copy to send holds the i1iSis and nothing else
    (no keyboard, no monitor calibrator). Without an i1iSis in it,
    only an X-Rite device the tool does not know is kept, and the
    tool says so. Without a USB description nothing is kept.
  • The analysis puts the steps side by side, says whether the
    i1iSis seems to talk through a USB-to-serial chip like the
    i1iO, and shows the first lines the program sent. The
    i1iSis's answers are never shown, only their sizes.
  • The read-only promise, tested as far as a test can: a test
    runs a whole simulated session while it watches and blocks the
    usual ways of opening files and starting programs (open,
    io.open, os.open, pathlib, subprocess, os.system,
    os.popen, os.posix_spawn, os.spawn*, os.exec*,
    os.fork, os.forkpty, the pty module, multiprocessing,
    and ctypes, which the tool may use only through Windows'
    kernel32 and shell32). It fails if the tool opens a device,
    writes outside its folders, or starts a program other than the
    ones it names with the arguments it really builds
    (PowerShell: the exact scripts); eighteen deliberately broken
    copies are all caught. A test cannot prove that nothing could
    slip through some other way.
  • Known limit: two plain files given to --analyze
    together that hold devices at the same USB address can be
    named after the wrong instrument family, and what the program
    sent to the other instrument can then ...
Read more

v1.6.4

Choose a tag to compare

@github-actions github-actions released this 14 Nov 12:45

Ready-to-run builds of the ChromIQ X-Rite instrument information
collector: for the i1iO table, since 1.5 for handheld instruments
(i1Pro, i1Pro 2, i1Pro 3 / 3 Plus, ColorMunki, i1Studio), and since
1.6 for the i1iSis chart readers (i1iSis, i1iSis XL, i1iSis 2,
i1iSis 2 XL).

Just want to run it? Download the file for your system:

System File
Windows chromiq-io-probe-windows.exe
macOS (Intel and Apple Silicon) chromiq-io-probe-macos
Linux chromiq-io-probe-linux

No Python needed. It changes nothing on the instrument or the
computer
: it never sends anything to the instrument or the table,
cannot move the i1iO's arm, and changes no settings. The optional
recording only listens (on Windows through the free USBPcap, which
you install yourself if you want to record).

What's new in 1.6.4

About older reports: reports made with a version before
1.6.3 contain codes that can be traced back to short serial
numbers. Please do not post such a report publicly, and delete
one you already posted if you can. --clean cannot change those
old codes.

  • Linux: a USB stick plugged into the instrument's old port
    later no longer shows its name in the report: a port counts
    as the instrument's only until it is unplugged. A kernel line
    that names no USB port is kept only when it carries the
    instrument's USB id, not because of a name in it.
  • Known instruments are named (for example a DTP20 or an
    i1iO 3 table) instead of "unknown X-Rite device". What the
    tool advises you to do is the same as before, except for an
    i1iO 3 table: like the other i1iO tables it now says to
    choose 1 (record an i1iO table) at the first question.
  • macOS: the few details kept for other USB devices (ids,
    version, speed) are kept only when they look like numbers of
    that kind; anything else is not shown.

What's new in 1.6.3

  • When a step keeps nothing: a tester's screenshot showed a
    Windows session in which every step stayed empty; the cause
    is not known
    . The tool still keeps nothing it cannot
    attribute. It now says what such a recording held (counts,
    kinds of device and USB ids only, never contents) and what to
    try, one action per line (replug the instrument; if still
    empty, restart Windows with "Restart", not "Shut down"). A
    session in which no step kept anything ends with one small
    SEND-THIS-<date>-report-only.zip: only the report and
    session.json, no recording.
  • A handheld session keeps only its own instrument family
    (the i1Pro and the i1Pro 2 count as one), as the i1iSis
    session does; an unknown X-Rite device only with a note.
  • The wrong instrument chosen: when a step holds only another
    instrument's traffic, the tool says so first, names the number
    to choose, and offers to stop (no replug advice then).
  • The report's device lists show other USB devices without
    their names
    (ids and kind only; an allow-list); --clean
    does the same for older reports; --no-redact shows
    everything and says so.
  • The short codes that replace serial numbers are salted per
    run
    and can no longer be traced back by trying numbers.
  • Steps shown after another device's traffic was removed print
    no decoded text; no decoded string carries control
    characters; 0xCF is shown as "clock mode N".
  • Analysis: ArgyllCMS does send the i1Pro 3's 0x3B (it was
    listed as never sent); the i1Pro 3's ruler travel per scan;
    every swipe by its length; the dual scan is not compared in
    "only with the ruler".
  • Steps: after the single patch, switch back to strip scanning;
    a lights answer "no" is asked once more.

What's new in 1.6.2

A privacy fix. Please read this if you made a recording with
1.6.1 or older.

  • Versions up to 1.6.1 could keep traffic of another USB
    device
    that shared the instrument's USB hub (for example a
    USB disk) in the files meant to hold only the instrument's
    traffic, and in analysis.txt and the report.
  • 1.6.2 leaves out traffic at the instrument's address on
    endpoints the instrument does not have (the case found), other
    devices' kinds of traffic (disk, keyboard, mouse, hub), and
    monitor calibrators. When another device was also seen on one
    of the instrument's own endpoints, the analysis shows no
    contents of that step and asks you to record it again.
  • It cannot tell apart a device that was given the
    instrument's address and uses the same endpoint numbers (for
    example a USB 3 "UAS" disk, a printer, a phone or a USB-serial
    adapter), nor another device's vendor requests.
  • What to do: update; clean old files with
    --clean SEND-THIS-....zip (or delete them); and while you
    record, unplug every other USB device you can, or at least
    those on the same hub or dock. If you already sent us an older
    file, tell us and we will delete it wherever we can, and treat
    it as private meanwhile.
  • Clearer steps: one action per line; wait for the ready signal
    before moving; "slow" means about twice as long as the normal
    scan, counted. The measurement condition is asked.
  • Analysis: lights in words, the ruler's motion per scan, fast
    and slow judged by the ruler.

What's new in 1.6.1

  • --analyze with several recordings of different
    instruments
    (for example an i1Pro 2 recording and an i1iSis
    recording) no longer reads them as one. It lists which
    instrument each file shows and asks you to analyse each
    instrument's recordings on their own. This closes the known
    limit of 1.6.0. Session folders, and what every session keeps
    in the copy to send, are exactly as in 1.6.0.
  • The i1iSis part of the README now says right there that the
    session has not been tried with a real i1iSis yet.

What's new in 1.6.0

The i1iSis chart readers. If you own an i1iSis, i1iSis XL,
i1iSis 2 or i1iSis 2 XL, the tool now offers a guided recording
session for it (or start it with --isis). It has not been
tried with a real i1iSis yet
: everything was tested with
invented recordings only, so your first session is the real
test. Nothing is promised: ChromIQ cannot use an i1iSis, and
whether it ever can is a separate question.

  • Six short steps in i1Profiler: start it and leave it idle;
    get the i1iSis ready for a page without feeding it; feed one
    page in single scan; the same page in dual scan; the same page
    a little crooked; a plain sheet that fails on purpose. After
    each step one extra question: how did the page move?
  • Switch the i1iSis on before plugging in its USB cable
    (X-Rite's own note). On Windows 11 25H2 use i1Profiler 3.8.8
    or newer, which X-Rite says fixed the i1iSis drivers there.
  • Any printed i1iSis chart page works (its colours do not
    matter for the recording); the tool says how to make one in
    i1Profiler and ships no chart.
  • Only the i1iSis is kept. When a recording shows the
    i1iSis, the copy to send holds the i1iSis and nothing else
    (no keyboard, no monitor calibrator). Without an i1iSis in it,
    only an X-Rite device the tool does not know is kept, and the
    tool says so. Without a USB description nothing is kept.
  • The analysis puts the steps side by side, says whether the
    i1iSis seems to talk through a USB-to-serial chip like the
    i1iO, and shows the first lines the program sent. The
    i1iSis's answers are never shown, only their sizes.
  • The read-only promise, tested as far as a test can: a test
    runs a whole simulated session while it watches and blocks the
    usual ways of opening files and starting programs (open,
    io.open, os.open, pathlib, subprocess, os.system,
    os.popen, os.posix_spawn, os.spawn*, os.exec*,
    os.fork, os.forkpty, the pty module, multiprocessing,
    and ctypes, which the tool may use only through Windows'
    kernel32 and shell32). It fails if the tool opens a device,
    writes outside its folders, or starts a program other than the
    ones it names with the arguments it really builds
    (PowerShell: the exact scripts); eighteen deliberately broken
    copies are all caught. A test cannot prove that nothing could
    slip through some other way.
  • Known limit: two plain files given to --analyze
    together that hold devices at the same USB address can be
    named after the wrong instrument family, and what the program
    sent to the other instrument can then appear among the lines
    sent to the i1iSis (the devices' answers stay hidden; the
    copies to send are not affected). Fixed in 1.6.1.
  • An independent review before the release found, and this
    release fixes: a device guessed from i1Pro commands could end
    up in the i1iSis copy when the recording had no USB
    description; the analysis of a raw recording could list a
    monitor calibrator as the instrument; a restart as
    administrator on Windows forgot --isis and --steps; some
    texts still spoke to i1Pro owners in an i1iSis session.
    A second review found that the first fixes made an i1iO table
    recording keep only an i1iSis on the same bus: now every
    session except the i1iSis keeps exactly what 1.5.5 kept,
    checked against the tagged 1.5.5 itself. A third review found
    that the analysis of a raw i1iSis recording with a monitor
    calibrator beside it mixed both; fixed, and the build's smoke
    steps are now also run locally before every release.

What's new in 1.5.5

Fixes from the first real Windows recording. Version 1.5.4
made a recording on Windows possible for the first time: a tester
recorded seven of nine steps with an i1Pro 3. Thank you! What his
session showed, and what 1.5.5 changes:

  • The Enter key: keys pressed while the tool is busy are now
    thrown away before every question (in 1.5.4 they answered the
    next questions, so two steps started and stopped at once). The
    tool says when it has your key and asks you to wait. A step
    stops only with e and Enter, and a step shorter than 2
    seconds asks whether that was on purpose.
  • **Nothing is...
Read more

v1.6.3

Choose a tag to compare

@github-actions github-actions released this 14 Nov 10:43

Ready-to-run builds of the ChromIQ X-Rite instrument information
collector: for the i1iO table, since 1.5 for handheld instruments
(i1Pro, i1Pro 2, i1Pro 3 / 3 Plus, ColorMunki, i1Studio), and since
1.6 for the i1iSis chart readers (i1iSis, i1iSis XL, i1iSis 2,
i1iSis 2 XL).

Just want to run it? Download the file for your system:

System File
Windows chromiq-io-probe-windows.exe
macOS (Intel and Apple Silicon) chromiq-io-probe-macos
Linux chromiq-io-probe-linux

No Python needed. It changes nothing on the instrument or the
computer
: it never sends anything to the instrument or the table,
cannot move the i1iO's arm, and changes no settings. The optional
recording only listens (on Windows through the free USBPcap, which
you install yourself if you want to record).

What's new in 1.6.3

Please read this if you have a report made with an older version:
Reports made with a version before 1.6.3 contain codes that can be traced back to short serial numbers. Please do not post such an older report publicly, and if you already posted one, delete it if you can. --clean cannot change those old codes.

  • Codes for serial numbers are now different on every run. The tool
    still replaces serial numbers with a short code, but that code is now
    mixed with a random value that is made new each time the tool runs and
    is never saved anywhere. Within one report the same serial number still
    gets the same code, but the code can no longer be traced back to the
    serial number, and it is not the same in two people's reports.
  • If you chose the wrong instrument at the start, the tool now tells you
    that first, right after the step: which instrument the recording shows,
    that nothing was kept, and which number to choose when you start the tool
    again. It offers to stop right there. The end of the run and the report
    say it again.
  • Other USB devices are hidden more strictly in the report's device
    lists on macOS, Linux and Windows. Another device keeps only its USB ids
    and its kind, not its name, a disk's volume name or its location. On Linux
    only the kernel messages about the X-Rite instrument are kept. --clean
    does the same for older reports.
  • --no-redact now warns you. It still shows everything, and the screen
    and the report now say: "Nothing is hidden in this report. Do not post it
    publicly."
  • When a step keeps nothing, the tool says what the recording held
    (counts and USB ids only, never contents) and what you can try, one step
    per line. Why every step of a session sometimes stays empty is not known
    yet. A session that kept nothing ends with one small
    SEND-THIS-<date>-report-only.zip with only the report and
    session.json, no recording.
  • A handheld session keeps only the instrument family you chose (the
    i1Pro and the i1Pro 2 count as one), never another instrument the tool
    knows.
  • Smaller things: clearer step texts after the single patch and for the
    lights question, and a more careful analysis (the i1Pro 3's ruler
    travel, every swipe by its length, decoded text never shown where
    another device's traffic was removed).

What's new in 1.6.2

A privacy fix. Please read this if you made a recording with
1.6.1 or older.

  • Versions up to 1.6.1 could keep traffic of another USB
    device
    that shared the instrument's USB hub (for example a
    USB disk) in the files meant to hold only the instrument's
    traffic, and in analysis.txt and the report.
  • 1.6.2 leaves out traffic at the instrument's address on
    endpoints the instrument does not have (the case found), other
    devices' kinds of traffic (disk, keyboard, mouse, hub), and
    monitor calibrators. When another device was also seen on one
    of the instrument's own endpoints, the analysis shows no
    contents of that step and asks you to record it again.
  • It cannot tell apart a device that was given the
    instrument's address and uses the same endpoint numbers (for
    example a USB 3 "UAS" disk, a printer, a phone or a USB-serial
    adapter), nor another device's vendor requests.
  • What to do: update; clean old files with
    --clean SEND-THIS-....zip (or delete them); and while you
    record, unplug every other USB device you can, or at least
    those on the same hub or dock. If you already sent us an older
    file, tell us and we will delete it wherever we can, and treat
    it as private meanwhile.
  • Clearer steps: one action per line; wait for the ready signal
    before moving; "slow" means about twice as long as the normal
    scan, counted. The measurement condition is asked.
  • Analysis: lights in words, the ruler's motion per scan, fast
    and slow judged by the ruler.

What's new in 1.6.1

  • --analyze with several recordings of different
    instruments
    (for example an i1Pro 2 recording and an i1iSis
    recording) no longer reads them as one. It lists which
    instrument each file shows and asks you to analyse each
    instrument's recordings on their own. This closes the known
    limit of 1.6.0. Session folders, and what every session keeps
    in the copy to send, are exactly as in 1.6.0.
  • The i1iSis part of the README now says right there that the
    session has not been tried with a real i1iSis yet.

What's new in 1.6.0

The i1iSis chart readers. If you own an i1iSis, i1iSis XL,
i1iSis 2 or i1iSis 2 XL, the tool now offers a guided recording
session for it (or start it with --isis). It has not been
tried with a real i1iSis yet
: everything was tested with
invented recordings only, so your first session is the real
test. Nothing is promised: ChromIQ cannot use an i1iSis, and
whether it ever can is a separate question.

  • Six short steps in i1Profiler: start it and leave it idle;
    get the i1iSis ready for a page without feeding it; feed one
    page in single scan; the same page in dual scan; the same page
    a little crooked; a plain sheet that fails on purpose. After
    each step one extra question: how did the page move?
  • Switch the i1iSis on before plugging in its USB cable
    (X-Rite's own note). On Windows 11 25H2 use i1Profiler 3.8.8
    or newer, which X-Rite says fixed the i1iSis drivers there.
  • Any printed i1iSis chart page works (its colours do not
    matter for the recording); the tool says how to make one in
    i1Profiler and ships no chart.
  • Only the i1iSis is kept. When a recording shows the
    i1iSis, the copy to send holds the i1iSis and nothing else
    (no keyboard, no monitor calibrator). Without an i1iSis in it,
    only an X-Rite device the tool does not know is kept, and the
    tool says so. Without a USB description nothing is kept.
  • The analysis puts the steps side by side, says whether the
    i1iSis seems to talk through a USB-to-serial chip like the
    i1iO, and shows the first lines the program sent. The
    i1iSis's answers are never shown, only their sizes.
  • The read-only promise, tested as far as a test can: a test
    runs a whole simulated session while it watches and blocks the
    usual ways of opening files and starting programs (open,
    io.open, os.open, pathlib, subprocess, os.system,
    os.popen, os.posix_spawn, os.spawn*, os.exec*,
    os.fork, os.forkpty, the pty module, multiprocessing,
    and ctypes, which the tool may use only through Windows'
    kernel32 and shell32). It fails if the tool opens a device,
    writes outside its folders, or starts a program other than the
    ones it names with the arguments it really builds
    (PowerShell: the exact scripts); eighteen deliberately broken
    copies are all caught. A test cannot prove that nothing could
    slip through some other way.
  • Known limit: two plain files given to --analyze
    together that hold devices at the same USB address can be
    named after the wrong instrument family, and what the program
    sent to the other instrument can then appear among the lines
    sent to the i1iSis (the devices' answers stay hidden; the
    copies to send are not affected). Fixed in 1.6.1.
  • An independent review before the release found, and this
    release fixes: a device guessed from i1Pro commands could end
    up in the i1iSis copy when the recording had no USB
    description; the analysis of a raw recording could list a
    monitor calibrator as the instrument; a restart as
    administrator on Windows forgot --isis and --steps; some
    texts still spoke to i1Pro owners in an i1iSis session.
    A second review found that the first fixes made an i1iO table
    recording keep only an i1iSis on the same bus: now every
    session except the i1iSis keeps exactly what 1.5.5 kept,
    checked against the tagged 1.5.5 itself. A third review found
    that the analysis of a raw i1iSis recording with a monitor
    calibrator beside it mixed both; fixed, and the build's smoke
    steps are now also run locally before every release.

What's new in 1.5.5

Fixes from the first real Windows recording. Version 1.5.4
made a recording on Windows possible for the first time: a tester
recorded seven of nine steps with an i1Pro 3. Thank you! What his
session showed, and what 1.5.5 changes:

  • The Enter key: keys pressed while the tool is busy are now
    thrown away before every question (in 1.5.4 they answered the
    next questions, so two steps started and stopped at once). The
    tool says when it has your key and asks you to wait. A step
    stops only with e and Enter, and a step shorter than 2
    seconds asks whether that was on purpose.
  • Nothing is drawn over what you type any more; the running
    count is in the window's title. In the classic Windows console
    the tool switches "QuickEdit" off during the session, so a click
    into the window cannot swallow the next Enter.
  • Empty steps are called EMPTY (please repeat), never "OK",
    and every handheld step gets a short check right after it.
  • Repeat only some steps: --steps 7,9, or the question at
    the start when an earlier session is next to t...
Read more

v1.6.2

Choose a tag to compare

@github-actions github-actions released this 13 Nov 23:04

Ready-to-run builds of the ChromIQ X-Rite instrument information
collector: for the i1iO table, since 1.5 for handheld instruments
(i1Pro, i1Pro 2, i1Pro 3 / 3 Plus, ColorMunki, i1Studio), and since
1.6 for the i1iSis chart readers (i1iSis, i1iSis XL, i1iSis 2,
i1iSis 2 XL).

Just want to run it? Download the file for your system:

System File
Windows chromiq-io-probe-windows.exe
macOS (Intel and Apple Silicon) chromiq-io-probe-macos
Linux chromiq-io-probe-linux

No Python needed. It changes nothing on the instrument or the
computer
: it never sends anything to the instrument or the table,
cannot move the i1iO's arm, and changes no settings. The optional
recording only listens (on Windows through the free USBPcap, which
you install yourself if you want to record).

What's new in 1.6.2

A privacy fix. Please read this if you made a recording with
1.6.1 or older.

  • Versions up to 1.6.1 could keep traffic of another USB
    device
    that shared the instrument's USB hub (for example a
    USB disk) in the files meant to hold only the instrument's
    traffic, and in analysis.txt and the report.
  • 1.6.2 leaves out traffic at the instrument's address on
    endpoints the instrument does not have (the case found), other
    devices' kinds of traffic (disk, keyboard, mouse, hub), and
    monitor calibrators. When another device was also seen on one
    of the instrument's own endpoints, the analysis shows no
    contents of that step and asks you to record it again.
  • It cannot tell apart a device that was given the
    instrument's address and uses the same endpoint numbers (for
    example a USB 3 "UAS" disk, a printer, a phone or a USB-serial
    adapter), nor another device's vendor requests.
  • What to do: update; clean old files with
    --clean SEND-THIS-....zip (or delete them); and while you
    record, unplug every other USB device you can, or at least
    those on the same hub or dock. If you already sent us an older
    file, tell us and we will delete it wherever we can, and treat
    it as private meanwhile.
  • Clearer steps: one action per line; wait for the ready signal
    before moving; "slow" means about twice as long as the normal
    scan, counted. The measurement condition is asked.
  • Analysis: lights in words, the ruler's motion per scan, fast
    and slow judged by the ruler.

What's new in 1.6.1

  • --analyze with several recordings of different
    instruments
    (for example an i1Pro 2 recording and an i1iSis
    recording) no longer reads them as one. It lists which
    instrument each file shows and asks you to analyse each
    instrument's recordings on their own. This closes the known
    limit of 1.6.0. Session folders, and what every session keeps
    in the copy to send, are exactly as in 1.6.0.
  • The i1iSis part of the README now says right there that the
    session has not been tried with a real i1iSis yet.

What's new in 1.6.0

The i1iSis chart readers. If you own an i1iSis, i1iSis XL,
i1iSis 2 or i1iSis 2 XL, the tool now offers a guided recording
session for it (or start it with --isis). It has not been
tried with a real i1iSis yet
: everything was tested with
invented recordings only, so your first session is the real
test. Nothing is promised: ChromIQ cannot use an i1iSis, and
whether it ever can is a separate question.

  • Six short steps in i1Profiler: start it and leave it idle;
    get the i1iSis ready for a page without feeding it; feed one
    page in single scan; the same page in dual scan; the same page
    a little crooked; a plain sheet that fails on purpose. After
    each step one extra question: how did the page move?
  • Switch the i1iSis on before plugging in its USB cable
    (X-Rite's own note). On Windows 11 25H2 use i1Profiler 3.8.8
    or newer, which X-Rite says fixed the i1iSis drivers there.
  • Any printed i1iSis chart page works (its colours do not
    matter for the recording); the tool says how to make one in
    i1Profiler and ships no chart.
  • Only the i1iSis is kept. When a recording shows the
    i1iSis, the copy to send holds the i1iSis and nothing else
    (no keyboard, no monitor calibrator). Without an i1iSis in it,
    only an X-Rite device the tool does not know is kept, and the
    tool says so. Without a USB description nothing is kept.
  • The analysis puts the steps side by side, says whether the
    i1iSis seems to talk through a USB-to-serial chip like the
    i1iO, and shows the first lines the program sent. The
    i1iSis's answers are never shown, only their sizes.
  • The read-only promise, tested as far as a test can: a test
    runs a whole simulated session while it watches and blocks the
    usual ways of opening files and starting programs (open,
    io.open, os.open, pathlib, subprocess, os.system,
    os.popen, os.posix_spawn, os.spawn*, os.exec*,
    os.fork, os.forkpty, the pty module, multiprocessing,
    and ctypes, which the tool may use only through Windows'
    kernel32 and shell32). It fails if the tool opens a device,
    writes outside its folders, or starts a program other than the
    ones it names with the arguments it really builds
    (PowerShell: the exact scripts); eighteen deliberately broken
    copies are all caught. A test cannot prove that nothing could
    slip through some other way.
  • Known limit: two plain files given to --analyze
    together that hold devices at the same USB address can be
    named after the wrong instrument family, and what the program
    sent to the other instrument can then appear among the lines
    sent to the i1iSis (the devices' answers stay hidden; the
    copies to send are not affected). Fixed in 1.6.1.
  • An independent review before the release found, and this
    release fixes: a device guessed from i1Pro commands could end
    up in the i1iSis copy when the recording had no USB
    description; the analysis of a raw recording could list a
    monitor calibrator as the instrument; a restart as
    administrator on Windows forgot --isis and --steps; some
    texts still spoke to i1Pro owners in an i1iSis session.
    A second review found that the first fixes made an i1iO table
    recording keep only an i1iSis on the same bus: now every
    session except the i1iSis keeps exactly what 1.5.5 kept,
    checked against the tagged 1.5.5 itself. A third review found
    that the analysis of a raw i1iSis recording with a monitor
    calibrator beside it mixed both; fixed, and the build's smoke
    steps are now also run locally before every release.

What's new in 1.5.5

Fixes from the first real Windows recording. Version 1.5.4
made a recording on Windows possible for the first time: a tester
recorded seven of nine steps with an i1Pro 3. Thank you! What his
session showed, and what 1.5.5 changes:

  • The Enter key: keys pressed while the tool is busy are now
    thrown away before every question (in 1.5.4 they answered the
    next questions, so two steps started and stopped at once). The
    tool says when it has your key and asks you to wait. A step
    stops only with e and Enter, and a step shorter than 2
    seconds asks whether that was on purpose.
  • Nothing is drawn over what you type any more; the running
    count is in the window's title. In the classic Windows console
    the tool switches "QuickEdit" off during the session, so a click
    into the window cannot swallow the next Enter.
  • Empty steps are called EMPTY (please repeat), never "OK",
    and every handheld step gets a short check right after it.
  • Repeat only some steps: --steps 7,9, or the question at
    the start when an earlier session is next to the tool.
  • Clearer idle, slow-swipe and dual-scan steps; the i1Pro 3's red
    light is named as red (likely).

What's new in 1.5.4

Aimed at the next recording of an i1iO table. The first real
recording of an i1iO 2 (with an i1Pro 2, on a Mac) now exists.
It showed that i1Profiler talks to the table in short lines of
text through a common USB-to-serial chip, and every one of its
2,475 lines passed the check number we worked out, so we very
likely read the format correctly. Much is still unknown, and
nothing is promised: ChromIQ cannot drive the table. There is
still no recording from a handheld instrument.

  • New i1iO steps, the most useful first: one row of a chart
    that i1Profiler scans in one pass, while you mark when the arm
    starts (a), the lamp lights (l) and the head goes down and up
    (d, u), then a cancel; a few patches in spot mode (if
    offered); a small chart to the end with the fastest setting; one
    row with dual scan (if offered); one row of the earlier chart (if
    at hand). Then a few questions (chart files, paper size, a ruler
    measurement, the settings, the sheet hold).
  • The i1iO 3 (0765:30B0) is recognised as a table; its protocol is
    not known yet.
  • On a Mac, every step is checked straight away ("this step
    recorded 14 commands to the table: good", or "please repeat
    it"), and a table at the end says what is still missing, all
    while SIP is still off. A list of what to have ready comes
    before the recording, and a box with the exact steps to switch
    SIP back on (and csrutil status to check) comes last.
  • --rehearse shows every step without recording anything.
    On a Mac: ./chromiq-io-probe-macos --rehearse in the Terminal.
  • Fixed on Windows: 1.5.3 could not record on Windows at all,
    because of our bug.
    The tool made the recording file before
    USBPcap could, and USBPcap refuses to write into a file that
    already exists, so the recording stopped at once every time
    (found thanks to a real report). 1.5.4 fixes this according to
    the recorder's own source code and our tests, but a real
    Windows recording has not confirmed it yet. If the start fails
    anyway, the report now holds everything needed to see why, and
    the tool offers to record by hand with Wireshark instead.
  • One file to send: at the end the tool makes
    SEND-THIS-<date>.zip next t...
Read more

v1.6.1

Choose a tag to compare

@github-actions github-actions released this 13 Nov 17:39

Ready-to-run builds of the ChromIQ X-Rite instrument information
collector: for the i1iO table, since 1.5 for handheld instruments
(i1Pro, i1Pro 2, i1Pro 3 / 3 Plus, ColorMunki, i1Studio), and since
1.6 for the i1iSis chart readers (i1iSis, i1iSis XL, i1iSis 2,
i1iSis 2 XL).

Just want to run it? Download the file for your system:

System File
Windows chromiq-io-probe-windows.exe
macOS (Intel and Apple Silicon) chromiq-io-probe-macos
Linux chromiq-io-probe-linux

No Python needed. It changes nothing on the instrument or the
computer
: it never sends anything to the instrument or the table,
cannot move the i1iO's arm, and changes no settings. The optional
recording only listens (on Windows through the free USBPcap, which
you install yourself if you want to record).

What's new in 1.6.1

  • --analyze with several recordings of different
    instruments
    (for example an i1Pro 2 recording and an i1iSis
    recording) no longer reads them as one. It lists which
    instrument each file shows and asks you to analyse each
    instrument's recordings on their own. This closes the known
    limit of 1.6.0. Session folders, and what every session keeps
    in the copy to send, are exactly as in 1.6.0.
  • The i1iSis part of the README now says right there that the
    session has not been tried with a real i1iSis yet.

What's new in 1.6.0

The i1iSis chart readers. If you own an i1iSis, i1iSis XL,
i1iSis 2 or i1iSis 2 XL, the tool now offers a guided recording
session for it (or start it with --isis). It has not been
tried with a real i1iSis yet
: everything was tested with
invented recordings only, so your first session is the real
test. Nothing is promised: ChromIQ cannot use an i1iSis, and
whether it ever can is a separate question.

  • Six short steps in i1Profiler: start it and leave it idle;
    get the i1iSis ready for a page without feeding it; feed one
    page in single scan; the same page in dual scan; the same page
    a little crooked; a plain sheet that fails on purpose. After
    each step one extra question: how did the page move?
  • Switch the i1iSis on before plugging in its USB cable
    (X-Rite's own note). On Windows 11 25H2 use i1Profiler 3.8.8
    or newer, which X-Rite says fixed the i1iSis drivers there.
  • Any printed i1iSis chart page works (its colours do not
    matter for the recording); the tool says how to make one in
    i1Profiler and ships no chart.
  • Only the i1iSis is kept. When a recording shows the
    i1iSis, the copy to send holds the i1iSis and nothing else
    (no keyboard, no monitor calibrator). Without an i1iSis in it,
    only an X-Rite device the tool does not know is kept, and the
    tool says so. Without a USB description nothing is kept.
  • The analysis puts the steps side by side, says whether the
    i1iSis seems to talk through a USB-to-serial chip like the
    i1iO, and shows the first lines the program sent. The
    i1iSis's answers are never shown, only their sizes.
  • The read-only promise, tested as far as a test can: a test
    runs a whole simulated session while it watches and blocks the
    usual ways of opening files and starting programs (open,
    io.open, os.open, pathlib, subprocess, os.system,
    os.popen, os.posix_spawn, os.spawn*, os.exec*,
    os.fork, os.forkpty, the pty module, multiprocessing,
    and ctypes, which the tool may use only through Windows'
    kernel32 and shell32). It fails if the tool opens a device,
    writes outside its folders, or starts a program other than the
    ones it names with the arguments it really builds
    (PowerShell: the exact scripts); eighteen deliberately broken
    copies are all caught. A test cannot prove that nothing could
    slip through some other way.
  • Known limit: two plain files given to --analyze
    together that hold devices at the same USB address can be
    named after the wrong instrument family, and what the program
    sent to the other instrument can then appear among the lines
    sent to the i1iSis (the devices' answers stay hidden; the
    copies to send are not affected). Fixed in 1.6.1.
  • An independent review before the release found, and this
    release fixes: a device guessed from i1Pro commands could end
    up in the i1iSis copy when the recording had no USB
    description; the analysis of a raw recording could list a
    monitor calibrator as the instrument; a restart as
    administrator on Windows forgot --isis and --steps; some
    texts still spoke to i1Pro owners in an i1iSis session.
    A second review found that the first fixes made an i1iO table
    recording keep only an i1iSis on the same bus: now every
    session except the i1iSis keeps exactly what 1.5.5 kept,
    checked against the tagged 1.5.5 itself. A third review found
    that the analysis of a raw i1iSis recording with a monitor
    calibrator beside it mixed both; fixed, and the build's smoke
    steps are now also run locally before every release.

What's new in 1.5.5

Fixes from the first real Windows recording. Version 1.5.4
made a recording on Windows possible for the first time: a tester
recorded seven of nine steps with an i1Pro 3. Thank you! What his
session showed, and what 1.5.5 changes:

  • The Enter key: keys pressed while the tool is busy are now
    thrown away before every question (in 1.5.4 they answered the
    next questions, so two steps started and stopped at once). The
    tool says when it has your key and asks you to wait. A step
    stops only with e and Enter, and a step shorter than 2
    seconds asks whether that was on purpose.
  • Nothing is drawn over what you type any more; the running
    count is in the window's title. In the classic Windows console
    the tool switches "QuickEdit" off during the session, so a click
    into the window cannot swallow the next Enter.
  • Empty steps are called EMPTY (please repeat), never "OK",
    and every handheld step gets a short check right after it.
  • Repeat only some steps: --steps 7,9, or the question at
    the start when an earlier session is next to the tool.
  • Clearer idle, slow-swipe and dual-scan steps; the i1Pro 3's red
    light is named as red (likely).

What's new in 1.5.4

Aimed at the next recording of an i1iO table. The first real
recording of an i1iO 2 (with an i1Pro 2, on a Mac) now exists.
It showed that i1Profiler talks to the table in short lines of
text through a common USB-to-serial chip, and every one of its
2,475 lines passed the check number we worked out, so we very
likely read the format correctly. Much is still unknown, and
nothing is promised: ChromIQ cannot drive the table. There is
still no recording from a handheld instrument.

  • New i1iO steps, the most useful first: one row of a chart
    that i1Profiler scans in one pass, while you mark when the arm
    starts (a), the lamp lights (l) and the head goes down and up
    (d, u), then a cancel; a few patches in spot mode (if
    offered); a small chart to the end with the fastest setting; one
    row with dual scan (if offered); one row of the earlier chart (if
    at hand). Then a few questions (chart files, paper size, a ruler
    measurement, the settings, the sheet hold).
  • The i1iO 3 (0765:30B0) is recognised as a table; its protocol is
    not known yet.
  • On a Mac, every step is checked straight away ("this step
    recorded 14 commands to the table: good", or "please repeat
    it"), and a table at the end says what is still missing, all
    while SIP is still off. A list of what to have ready comes
    before the recording, and a box with the exact steps to switch
    SIP back on (and csrutil status to check) comes last.
  • --rehearse shows every step without recording anything.
    On a Mac: ./chromiq-io-probe-macos --rehearse in the Terminal.
  • Fixed on Windows: 1.5.3 could not record on Windows at all,
    because of our bug.
    The tool made the recording file before
    USBPcap could, and USBPcap refuses to write into a file that
    already exists, so the recording stopped at once every time
    (found thanks to a real report). 1.5.4 fixes this according to
    the recorder's own source code and our tests, but a real
    Windows recording has not confirmed it yet. If the start fails
    anyway, the report now holds everything needed to see why, and
    the tool offers to record by hand with Wireshark instead.
  • One file to send: at the end the tool makes
    SEND-THIS-<date>.zip next to itself (the report and the
    instrument-only recordings), and a report without a recording
    says so at the top. Each handheld step is now a small framed
    list (Enter here, then the program, then Enter here again), with
    a line that shows the recording running.
  • Fixed: on a Mac, --analyze left out everything the program
    sent to a device (it sits in a different part of a macOS
    recording). Found in the first real recording.
  • --analyze reads the i1iO table's conversation: every line,
    its check number, and what it most likely means, each with
    "certain", "likely" or "guess". The serial number is hidden.
  • Windows on ARM and the driver help are worded only as far as
    the evidence goes (nothing is claimed about the i1Pro 3 on ARM
    that nobody has tried; no unverified Device Manager names).
  • An independent review before the release found, and this
    release fixes: stray keys could end or start a step, a second
    Ctrl-C at the end could lose the step log, a stopped recorder
    was reported as "please repeat", and the SIP box also came when
    nothing was recorded.

Tested: on a Mac with SIP on (recording not possible, as it
should be), the timed steps, the rehearsal and the analysis, in a
terminal of 80 columns; the analysis against the two real i1iO
recordings (the same 1,125 and 1,350 lines as our separate
decoder, every check number right). Not tested: a recording with
SIP off, and a real table (we have none).

What's new in 1.5.3

Help with the step where both users so far got stuck, and one
pl...

Read more

v1.6.0

Choose a tag to compare

@github-actions github-actions released this 13 Nov 17:07

Ready-to-run builds of the ChromIQ X-Rite instrument information
collector: for the i1iO table, since 1.5 for handheld instruments
(i1Pro, i1Pro 2, i1Pro 3 / 3 Plus, ColorMunki, i1Studio), and since
1.6 for the i1iSis chart readers (i1iSis, i1iSis XL, i1iSis 2,
i1iSis 2 XL).

Just want to run it? Download the file for your system:

System File
Windows chromiq-io-probe-windows.exe
macOS (Intel and Apple Silicon) chromiq-io-probe-macos
Linux chromiq-io-probe-linux

No Python needed. It changes nothing on the instrument or the
computer
: it never sends anything to the instrument or the table,
cannot move the i1iO's arm, and changes no settings. The optional
recording only listens (on Windows through the free USBPcap, which
you install yourself if you want to record).

What's new in 1.6.0

The i1iSis chart readers. If you own an i1iSis, i1iSis XL,
i1iSis 2 or i1iSis 2 XL, the tool now offers a guided recording
session for it (or start it with --isis). It has not been
tried with a real i1iSis yet
: everything was tested with
invented recordings only, so your first session is the real
test. Nothing is promised: ChromIQ cannot use an i1iSis, and
whether it ever can is a separate question.

  • Six short steps in i1Profiler: start it and leave it idle;
    get the i1iSis ready for a page without feeding it; feed one
    page in single scan; the same page in dual scan; the same page
    a little crooked; a plain sheet that fails on purpose. After
    each step one extra question: how did the page move?
  • Switch the i1iSis on before plugging in its USB cable
    (X-Rite's own note). On Windows 11 25H2 use i1Profiler 3.8.8
    or newer, which X-Rite says fixed the i1iSis drivers there.
  • Any printed i1iSis chart page works (its colours do not
    matter for the recording); the tool says how to make one in
    i1Profiler and ships no chart.
  • Only the i1iSis is kept. When a recording shows the
    i1iSis, the copy to send holds the i1iSis and nothing else
    (no keyboard, no monitor calibrator). Without an i1iSis in it,
    only an X-Rite device the tool does not know is kept, and the
    tool says so. Without a USB description nothing is kept.
  • The analysis puts the steps side by side, says whether the
    i1iSis seems to talk through a USB-to-serial chip like the
    i1iO, and shows the first lines the program sent. The
    i1iSis's answers are never shown, only their sizes.
  • The read-only promise, tested as far as a test can: a test
    runs a whole simulated session while it watches and blocks the
    usual ways of opening files and starting programs (open,
    io.open, os.open, pathlib, subprocess, os.system,
    os.popen, os.posix_spawn, os.spawn*, os.exec*,
    os.fork, os.forkpty, the pty module, multiprocessing,
    and ctypes, which the tool may use only through Windows'
    kernel32 and shell32). It fails if the tool opens a device,
    writes outside its folders, or starts a program other than the
    ones it names with the arguments it really builds
    (PowerShell: the exact scripts); eighteen deliberately broken
    copies are all caught. A test cannot prove that nothing could
    slip through some other way.
  • Known limit: two plain files given to --analyze
    together that hold devices at the same USB address can be
    named after the wrong instrument family, and what the program
    sent to the other instrument can then appear among the lines
    sent to the i1iSis (the devices' answers stay hidden; the
    copies to send are not affected). Analyse one session folder
    at a time; a later version fixes it.
  • An independent review before the release found, and this
    release fixes: a device guessed from i1Pro commands could end
    up in the i1iSis copy when the recording had no USB
    description; the analysis of a raw recording could list a
    monitor calibrator as the instrument; a restart as
    administrator on Windows forgot --isis and --steps; some
    texts still spoke to i1Pro owners in an i1iSis session.
    A second review found that the first fixes made an i1iO table
    recording keep only an i1iSis on the same bus: now every
    session except the i1iSis keeps exactly what 1.5.5 kept,
    checked against the tagged 1.5.5 itself. A third review found
    that the analysis of a raw i1iSis recording with a monitor
    calibrator beside it mixed both; fixed, and the build's smoke
    steps are now also run locally before every release.

What's new in 1.5.5

Fixes from the first real Windows recording. Version 1.5.4
made a recording on Windows possible for the first time: a tester
recorded seven of nine steps with an i1Pro 3. Thank you! What his
session showed, and what 1.5.5 changes:

  • The Enter key: keys pressed while the tool is busy are now
    thrown away before every question (in 1.5.4 they answered the
    next questions, so two steps started and stopped at once). The
    tool says when it has your key and asks you to wait. A step
    stops only with e and Enter, and a step shorter than 2
    seconds asks whether that was on purpose.
  • Nothing is drawn over what you type any more; the running
    count is in the window's title. In the classic Windows console
    the tool switches "QuickEdit" off during the session, so a click
    into the window cannot swallow the next Enter.
  • Empty steps are called EMPTY (please repeat), never "OK",
    and every handheld step gets a short check right after it.
  • Repeat only some steps: --steps 7,9, or the question at
    the start when an earlier session is next to the tool.
  • Clearer idle, slow-swipe and dual-scan steps; the i1Pro 3's red
    light is named as red (likely).

What's new in 1.5.4

Aimed at the next recording of an i1iO table. The first real
recording of an i1iO 2 (with an i1Pro 2, on a Mac) now exists.
It showed that i1Profiler talks to the table in short lines of
text through a common USB-to-serial chip, and every one of its
2,475 lines passed the check number we worked out, so we very
likely read the format correctly. Much is still unknown, and
nothing is promised: ChromIQ cannot drive the table. There is
still no recording from a handheld instrument.

  • New i1iO steps, the most useful first: one row of a chart
    that i1Profiler scans in one pass, while you mark when the arm
    starts (a), the lamp lights (l) and the head goes down and up
    (d, u), then a cancel; a few patches in spot mode (if
    offered); a small chart to the end with the fastest setting; one
    row with dual scan (if offered); one row of the earlier chart (if
    at hand). Then a few questions (chart files, paper size, a ruler
    measurement, the settings, the sheet hold).
  • The i1iO 3 (0765:30B0) is recognised as a table; its protocol is
    not known yet.
  • On a Mac, every step is checked straight away ("this step
    recorded 14 commands to the table: good", or "please repeat
    it"), and a table at the end says what is still missing, all
    while SIP is still off. A list of what to have ready comes
    before the recording, and a box with the exact steps to switch
    SIP back on (and csrutil status to check) comes last.
  • --rehearse shows every step without recording anything.
    On a Mac: ./chromiq-io-probe-macos --rehearse in the Terminal.
  • Fixed on Windows: 1.5.3 could not record on Windows at all,
    because of our bug.
    The tool made the recording file before
    USBPcap could, and USBPcap refuses to write into a file that
    already exists, so the recording stopped at once every time
    (found thanks to a real report). 1.5.4 fixes this according to
    the recorder's own source code and our tests, but a real
    Windows recording has not confirmed it yet. If the start fails
    anyway, the report now holds everything needed to see why, and
    the tool offers to record by hand with Wireshark instead.
  • One file to send: at the end the tool makes
    SEND-THIS-<date>.zip next to itself (the report and the
    instrument-only recordings), and a report without a recording
    says so at the top. Each handheld step is now a small framed
    list (Enter here, then the program, then Enter here again), with
    a line that shows the recording running.
  • Fixed: on a Mac, --analyze left out everything the program
    sent to a device (it sits in a different part of a macOS
    recording). Found in the first real recording.
  • --analyze reads the i1iO table's conversation: every line,
    its check number, and what it most likely means, each with
    "certain", "likely" or "guess". The serial number is hidden.
  • Windows on ARM and the driver help are worded only as far as
    the evidence goes (nothing is claimed about the i1Pro 3 on ARM
    that nobody has tried; no unverified Device Manager names).
  • An independent review before the release found, and this
    release fixes: stray keys could end or start a step, a second
    Ctrl-C at the end could lose the step log, a stopped recorder
    was reported as "please repeat", and the SIP box also came when
    nothing was recorded.

Tested: on a Mac with SIP on (recording not possible, as it
should be), the timed steps, the rehearsal and the analysis, in a
terminal of 80 columns; the analysis against the two real i1iO
recordings (the same 1,125 and 1,350 lines as our separate
decoder, every check number right). Not tested: a recording with
SIP off, and a real table (we have none).

What's new in 1.5.3

Help with the step where both users so far got stuck, and one
plain question instead of guessing what you want to do. Still
untested with a real instrument and a real recording.

  • "What would you like to do?" After looking for your
    instruments the tool shows what is plugged in right now (with
    its driver on Windows) and asks: 1 record an i1iO table, 2
    record a handheld instrument, 3 just make the report, or a to
    look again after plugging something in. The suggestion comes
    only from what is plugged in now, so Enter is usually right.
    --io an...
Read more

v1.5.5

Choose a tag to compare

@github-actions github-actions released this 13 Nov 06:11

Ready-to-run builds of the ChromIQ X-Rite instrument information
collector: for the i1iO table and, since 1.5, for handheld
instruments (i1Pro, i1Pro 2, i1Pro 3 / 3 Plus, ColorMunki, i1Studio).

Just want to run it? Download the file for your system:

System File
Windows chromiq-io-probe-windows.exe
macOS (Intel and Apple Silicon) chromiq-io-probe-macos
Linux chromiq-io-probe-linux

No Python needed. It changes nothing on the instrument or the
computer
: it never sends anything to the instrument or the table,
cannot move the i1iO's arm, and changes no settings. The optional
recording only listens (on Windows through the free USBPcap, which
you install yourself if you want to record).

What's new in 1.5.5

Fixes from the first real Windows recording. Version 1.5.4
made a recording on Windows possible for the first time: a tester
recorded seven of nine steps with an i1Pro 3. Thank you! What his
session showed, and what 1.5.5 changes:

  • The Enter key: keys pressed while the tool is busy are now
    thrown away before every question (in 1.5.4 they answered the
    next questions, so two steps started and stopped at once). The
    tool says when it has your key and asks you to wait. A step
    stops only with e and Enter, and a step shorter than 2
    seconds asks whether that was on purpose.
  • Nothing is drawn over what you type any more; the running
    count is in the window's title. In the classic Windows console
    the tool switches "QuickEdit" off during the session, so a click
    into the window cannot swallow the next Enter.
  • Empty steps are called EMPTY (please repeat), never "OK",
    and every handheld step gets a short check right after it.
  • Repeat only some steps: --steps 7,9, or the question at
    the start when an earlier session is next to the tool.
  • Clearer idle, slow-swipe and dual-scan steps; the i1Pro 3's red
    light is named as red (likely).

What's new in 1.5.4

Aimed at the next recording of an i1iO table. The first real
recording of an i1iO 2 (with an i1Pro 2, on a Mac) now exists.
It showed that i1Profiler talks to the table in short lines of
text through a common USB-to-serial chip, and every one of its
2,475 lines passed the check number we worked out, so we very
likely read the format correctly. Much is still unknown, and
nothing is promised: ChromIQ cannot drive the table. There is
still no recording from a handheld instrument.

  • New i1iO steps, the most useful first: one row of a chart
    that i1Profiler scans in one pass, while you mark when the arm
    starts (a), the lamp lights (l) and the head goes down and up
    (d, u), then a cancel; a few patches in spot mode (if
    offered); a small chart to the end with the fastest setting; one
    row with dual scan (if offered); one row of the earlier chart (if
    at hand). Then a few questions (chart files, paper size, a ruler
    measurement, the settings, the sheet hold).
  • The i1iO 3 (0765:30B0) is recognised as a table; its protocol is
    not known yet.
  • On a Mac, every step is checked straight away ("this step
    recorded 14 commands to the table: good", or "please repeat
    it"), and a table at the end says what is still missing, all
    while SIP is still off. A list of what to have ready comes
    before the recording, and a box with the exact steps to switch
    SIP back on (and csrutil status to check) comes last.
  • --rehearse shows every step without recording anything.
    On a Mac: ./chromiq-io-probe-macos --rehearse in the Terminal.
  • Fixed on Windows: 1.5.3 could not record on Windows at all,
    because of our bug.
    The tool made the recording file before
    USBPcap could, and USBPcap refuses to write into a file that
    already exists, so the recording stopped at once every time
    (found thanks to a real report). 1.5.4 fixes this according to
    the recorder's own source code and our tests, but a real
    Windows recording has not confirmed it yet. If the start fails
    anyway, the report now holds everything needed to see why, and
    the tool offers to record by hand with Wireshark instead.
  • One file to send: at the end the tool makes
    SEND-THIS-<date>.zip next to itself (the report and the
    instrument-only recordings), and a report without a recording
    says so at the top. Each handheld step is now a small framed
    list (Enter here, then the program, then Enter here again), with
    a line that shows the recording running.
  • Fixed: on a Mac, --analyze left out everything the program
    sent to a device (it sits in a different part of a macOS
    recording). Found in the first real recording.
  • --analyze reads the i1iO table's conversation: every line,
    its check number, and what it most likely means, each with
    "certain", "likely" or "guess". The serial number is hidden.
  • Windows on ARM and the driver help are worded only as far as
    the evidence goes (nothing is claimed about the i1Pro 3 on ARM
    that nobody has tried; no unverified Device Manager names).
  • An independent review before the release found, and this
    release fixes: stray keys could end or start a step, a second
    Ctrl-C at the end could lose the step log, a stopped recorder
    was reported as "please repeat", and the SIP box also came when
    nothing was recorded.

Tested: on a Mac with SIP on (recording not possible, as it
should be), the timed steps, the rehearsal and the analysis, in a
terminal of 80 columns; the analysis against the two real i1iO
recordings (the same 1,125 and 1,350 lines as our separate
decoder, every check number right). Not tested: a recording with
SIP off, and a real table (we have none).

What's new in 1.5.3

Help with the step where both users so far got stuck, and one
plain question instead of guessing what you want to do. Still
untested with a real instrument and a real recording.

  • "What would you like to do?" After looking for your
    instruments the tool shows what is plugged in right now (with
    its driver on Windows) and asks: 1 record an i1iO table, 2
    record a handheld instrument, 3 just make the report, or a to
    look again after plugging something in. The suggestion comes
    only from what is plugged in now, so Enter is usually right.
    --io and --i1pro choose without asking.
  • USBPcap missing: the tool can fetch and start its official
    installer, only after a yes.
    It first says what USBPcap is
    (free, open-source, by its own author, not by us), that it
    installs a system driver, that one user reported a blue screen
    naming its driver on Windows 11 25H2 (USBPcap issue #146), and
    how to remove it. The download (USBPcapSetup-1.5.4.0.exe from
    USBPcap's GitHub release) goes into a new private folder and is
    checked against a SHA-256 fingerprint written into the tool,
    after the download and again just before the start; on any
    difference it is deleted and never started. A second yes starts
    the installer the normal, visible way, with Windows' own
    permission question. The tool never restarts the computer.
  • Not administrator: one offer to start the tool again as
    administrator through Windows' own question.
  • Every "no" or failure leads to the manual steps, as in 1.5.2.
  • Helper programs never read your keys or write into the window;
    a device Windows only remembers is no longer offered as
    plugged in; the "remembered" note no longer wraps at 80 columns.
  • An independent review before the release found, and this
    release fixes: a half-written installer could stay behind; the
    install steps were repeated after a successful install; Ctrl-C
    at the question took the suggestion instead of stopping; the
    installer was saved next to the tool (now a private folder).

What was tested for real, and what was not

For the first time the Windows exe was driven on a real Windows:
Windows 11 Home 25H2 on ARM in a virtual machine (the x64 exe runs
emulated), German, by an automated test (tools/vm_test/) plus
two steps where Windows' permission window had to be answered:

  • double-click start; the window stays open at every ending
    (normal end, "just the report", choices 1 and 2, an error,
    a real Ctrl-C while it looks for devices) and closes on Enter; run from a console with output
    redirected it never waits;
  • the new question with nothing plugged in, "look again",
    choices 1, 2 and 3; an 80-column window; no garbled characters;
  • not administrator: the explanation before the steps, the offer,
    the real Windows permission window, and the copy started as
    administrator running;
  • USBPcap missing: the explanation, the real download from GitHub
    inside the virtual machine, a deliberately wrong fingerprint
    refused, the real permission window for USBPcap's installer
    (verified publisher "Tomasz Moń"), its licence page, and, after
    it was cancelled, the tool saying plainly that USBPcap is still
    not installed and cleaning up its download folder.

Not tested: a real instrument, a real recording, USBPcap's driver
actually installed (there is no USBPcap driver for Windows on ARM,
so it was not installed), and an x64 Windows PC.

What's new in 1.5.2

Changed after the second report from a user: the window closed
before the instructions could be read. Still untested with a real
recording.

  • The window stays open. Double-clicked on Windows, the window
    closed the moment the tool finished, so its help vanished and it
    looked like a crash. Every ending (normal end, nothing recorded,
    q, Ctrl-C, an error) now finishes with "Press Enter to close this
    window." It waits only in a console window of its own; from a
    command window, piped, redirected or under CI it does not.
    --pause and --no-pause force it either way.
  • WHAT TO DO NOW: when USBPcap, administrator rights, X-Rite's
    driver or the instrument is missing, a short box with the next
    steps is the last thing on the screen, with the name and folder
    of the report that holds the full instructions. When USBPca...
Read more

v1.5.4

Choose a tag to compare

@github-actions github-actions released this 13 Nov 02:56

Ready-to-run builds of the ChromIQ X-Rite instrument information
collector: for the i1iO table and, since 1.5, for handheld
instruments (i1Pro, i1Pro 2, i1Pro 3 / 3 Plus, ColorMunki, i1Studio).

Just want to run it? Download the file for your system:

System File
Windows chromiq-io-probe-windows.exe
macOS (Intel and Apple Silicon) chromiq-io-probe-macos
Linux chromiq-io-probe-linux

No Python needed. It changes nothing on the instrument or the
computer
: it never sends anything to the instrument or the table,
cannot move the i1iO's arm, and changes no settings. The optional
recording only listens (on Windows through the free USBPcap, which
you install yourself if you want to record).

What's new in 1.5.4

Aimed at the next recording of an i1iO table. The first real
recording of an i1iO 2 (with an i1Pro 2, on a Mac) now exists.
It showed that i1Profiler talks to the table in short lines of
text through a common USB-to-serial chip, and every one of its
2,475 lines passed the check number we worked out, so we very
likely read the format correctly. Much is still unknown, and
nothing is promised: ChromIQ cannot drive the table. There is
still no recording from a handheld instrument.

  • New i1iO steps, the most useful first: one row of a chart
    that i1Profiler scans in one pass, while you mark when the arm
    starts (a), the lamp lights (l) and the head goes down and up
    (d, u), then a cancel; a few patches in spot mode (if
    offered); a small chart to the end with the fastest setting; one
    row with dual scan (if offered); one row of the earlier chart (if
    at hand). Then a few questions (chart files, paper size, a ruler
    measurement, the settings, the sheet hold).
  • The i1iO 3 (0765:30B0) is recognised as a table; its protocol is
    not known yet.
  • On a Mac, every step is checked straight away ("this step
    recorded 14 commands to the table: good", or "please repeat
    it"), and a table at the end says what is still missing, all
    while SIP is still off. A list of what to have ready comes
    before the recording, and a box with the exact steps to switch
    SIP back on (and csrutil status to check) comes last.
  • --rehearse shows every step without recording anything.
    On a Mac: ./chromiq-io-probe-macos --rehearse in the Terminal.
  • Fixed on Windows: 1.5.3 could not record on Windows at all,
    because of our bug.
    The tool made the recording file before
    USBPcap could, and USBPcap refuses to write into a file that
    already exists, so the recording stopped at once every time
    (found thanks to a real report). 1.5.4 fixes this according to
    the recorder's own source code and our tests, but a real
    Windows recording has not confirmed it yet. If the start fails
    anyway, the report now holds everything needed to see why, and
    the tool offers to record by hand with Wireshark instead.
  • One file to send: at the end the tool makes
    SEND-THIS-<date>.zip next to itself (the report and the
    instrument-only recordings), and a report without a recording
    says so at the top. Each handheld step is now a small framed
    list (Enter here, then the program, then Enter here again), with
    a line that shows the recording running.
  • Fixed: on a Mac, --analyze left out everything the program
    sent to a device (it sits in a different part of a macOS
    recording). Found in the first real recording.
  • --analyze reads the i1iO table's conversation: every line,
    its check number, and what it most likely means, each with
    "certain", "likely" or "guess". The serial number is hidden.
  • Windows on ARM and the driver help are worded only as far as
    the evidence goes (nothing is claimed about the i1Pro 3 on ARM
    that nobody has tried; no unverified Device Manager names).
  • An independent review before the release found, and this
    release fixes: stray keys could end or start a step, a second
    Ctrl-C at the end could lose the step log, a stopped recorder
    was reported as "please repeat", and the SIP box also came when
    nothing was recorded.

Tested: on a Mac with SIP on (recording not possible, as it
should be), the timed steps, the rehearsal and the analysis, in a
terminal of 80 columns; the analysis against the two real i1iO
recordings (the same 1,125 and 1,350 lines as our separate
decoder, every check number right). Not tested: a recording with
SIP off, and a real table (we have none).

What's new in 1.5.3

Help with the step where both users so far got stuck, and one
plain question instead of guessing what you want to do. Still
untested with a real instrument and a real recording.

  • "What would you like to do?" After looking for your
    instruments the tool shows what is plugged in right now (with
    its driver on Windows) and asks: 1 record an i1iO table, 2
    record a handheld instrument, 3 just make the report, or a to
    look again after plugging something in. The suggestion comes
    only from what is plugged in now, so Enter is usually right.
    --io and --i1pro choose without asking.
  • USBPcap missing: the tool can fetch and start its official
    installer, only after a yes.
    It first says what USBPcap is
    (free, open-source, by its own author, not by us), that it
    installs a system driver, that one user reported a blue screen
    naming its driver on Windows 11 25H2 (USBPcap issue #146), and
    how to remove it. The download (USBPcapSetup-1.5.4.0.exe from
    USBPcap's GitHub release) goes into a new private folder and is
    checked against a SHA-256 fingerprint written into the tool,
    after the download and again just before the start; on any
    difference it is deleted and never started. A second yes starts
    the installer the normal, visible way, with Windows' own
    permission question. The tool never restarts the computer.
  • Not administrator: one offer to start the tool again as
    administrator through Windows' own question.
  • Every "no" or failure leads to the manual steps, as in 1.5.2.
  • Helper programs never read your keys or write into the window;
    a device Windows only remembers is no longer offered as
    plugged in; the "remembered" note no longer wraps at 80 columns.
  • An independent review before the release found, and this
    release fixes: a half-written installer could stay behind; the
    install steps were repeated after a successful install; Ctrl-C
    at the question took the suggestion instead of stopping; the
    installer was saved next to the tool (now a private folder).

What was tested for real, and what was not

For the first time the Windows exe was driven on a real Windows:
Windows 11 Home 25H2 on ARM in a virtual machine (the x64 exe runs
emulated), German, by an automated test (tools/vm_test/) plus
two steps where Windows' permission window had to be answered:

  • double-click start; the window stays open at every ending
    (normal end, "just the report", choices 1 and 2, an error,
    a real Ctrl-C while it looks for devices) and closes on Enter; run from a console with output
    redirected it never waits;
  • the new question with nothing plugged in, "look again",
    choices 1, 2 and 3; an 80-column window; no garbled characters;
  • not administrator: the explanation before the steps, the offer,
    the real Windows permission window, and the copy started as
    administrator running;
  • USBPcap missing: the explanation, the real download from GitHub
    inside the virtual machine, a deliberately wrong fingerprint
    refused, the real permission window for USBPcap's installer
    (verified publisher "Tomasz Moń"), its licence page, and, after
    it was cancelled, the tool saying plainly that USBPcap is still
    not installed and cleaning up its download folder.

Not tested: a real instrument, a real recording, USBPcap's driver
actually installed (there is no USBPcap driver for Windows on ARM,
so it was not installed), and an x64 Windows PC.

What's new in 1.5.2

Changed after the second report from a user: the window closed
before the instructions could be read. Still untested with a real
recording.

  • The window stays open. Double-clicked on Windows, the window
    closed the moment the tool finished, so its help vanished and it
    looked like a crash. Every ending (normal end, nothing recorded,
    q, Ctrl-C, an error) now finishes with "Press Enter to close this
    window." It waits only in a console window of its own; from a
    command window, piped, redirected or under CI it does not.
    --pause and --no-pause force it either way.
  • WHAT TO DO NOW: when USBPcap, administrator rights, X-Rite's
    driver or the instrument is missing, a short box with the next
    steps is the last thing on the screen, with the name and folder
    of the report that holds the full instructions. When USBPcap is
    missing it offers to open Wireshark's download page (only on a
    yes).
  • Administrator rights are checked before the steps: USBPcap
    lets only administrators listen to the USB ports, so the tool now
    says so first and explains how to start it as administrator,
    instead of letting the session begin.
  • An unexpected error says sorry, saves what was collected with
    the technical details, says where, and waits.
  • i1Pro 2 on Windows 11 25H2 or newer: when X-Rite's driver is
    installed and Windows reports the instrument as working, the tool
    says so and asks to carry on and tell us whether i1Profiler
    measures with it.

What's new in 1.5.1

Changed after the first report from a user. Still untested with a
real recording.

  • Clearer help when USBPcap is missing: what it is, why it is
    needed, numbered install steps (Wireshark's "USB Capture" page,
    the restart, run the tool again), and that no security setting
    has to be switched off. When nothing could be recorded, the run
    now ends by saying so, with what to do, instead of "ALL DONE".
  • The Windows 11 25H2 driver note only where it applies: X-Rite
    says the i1Pro 2 / iO2 driver is blocked on 25H2 and newer...
Read more

v1.5.3

Choose a tag to compare

@github-actions github-actions released this 12 Nov 22:07

Ready-to-run builds of the ChromIQ X-Rite instrument information
collector: for the i1iO table and, since 1.5, for handheld
instruments (i1Pro, i1Pro 2, i1Pro 3 / 3 Plus, ColorMunki, i1Studio).

Just want to run it? Download the file for your system:

System File
Windows chromiq-io-probe-windows.exe
macOS (Intel and Apple Silicon) chromiq-io-probe-macos
Linux chromiq-io-probe-linux

No Python needed. It changes nothing on the instrument or the
computer
: it never sends anything to the instrument or the table,
cannot move the i1iO's arm, and changes no settings. The optional
recording only listens (on Windows through the free USBPcap, which
you install yourself if you want to record).

What's new in 1.5.3

Help with the step where both users so far got stuck, and one
plain question instead of guessing what you want to do. Still
untested with a real instrument and a real recording.

  • "What would you like to do?" After looking for your
    instruments the tool shows what is plugged in right now (with
    its driver on Windows) and asks: 1 record an i1iO table, 2
    record a handheld instrument, 3 just make the report, or a to
    look again after plugging something in. The suggestion comes
    only from what is plugged in now, so Enter is usually right.
    --io and --i1pro choose without asking.
  • USBPcap missing: the tool can fetch and start its official
    installer, only after a yes.
    It first says what USBPcap is
    (free, open-source, by its own author, not by us), that it
    installs a system driver, that one user reported a blue screen
    naming its driver on Windows 11 25H2 (USBPcap issue #146), and
    how to remove it. The download (USBPcapSetup-1.5.4.0.exe from
    USBPcap's GitHub release) goes into a new private folder and is
    checked against a SHA-256 fingerprint written into the tool,
    after the download and again just before the start; on any
    difference it is deleted and never started. A second yes starts
    the installer the normal, visible way, with Windows' own
    permission question. The tool never restarts the computer.
  • Not administrator: one offer to start the tool again as
    administrator through Windows' own question.
  • Every "no" or failure leads to the manual steps, as in 1.5.2.
  • Helper programs never read your keys or write into the window;
    a device Windows only remembers is no longer offered as
    plugged in; the "remembered" note no longer wraps at 80 columns.
  • An independent review before the release found, and this
    release fixes: a half-written installer could stay behind; the
    install steps were repeated after a successful install; Ctrl-C
    at the question took the suggestion instead of stopping; the
    installer was saved next to the tool (now a private folder).

What was tested for real, and what was not

For the first time the Windows exe was driven on a real Windows:
Windows 11 Home 25H2 on ARM in a virtual machine (the x64 exe runs
emulated), German, by an automated test (tools/vm_test/) plus
two steps where Windows' permission window had to be answered:

  • double-click start; the window stays open at every ending
    (normal end, "just the report", choices 1 and 2, an error,
    a real Ctrl-C while it looks for devices) and closes on Enter; run from a console with output
    redirected it never waits;
  • the new question with nothing plugged in, "look again",
    choices 1, 2 and 3; an 80-column window; no garbled characters;
  • not administrator: the explanation before the steps, the offer,
    the real Windows permission window, and the copy started as
    administrator running;
  • USBPcap missing: the explanation, the real download from GitHub
    inside the virtual machine, a deliberately wrong fingerprint
    refused, the real permission window for USBPcap's installer
    (verified publisher "Tomasz Moń"), its licence page, and, after
    it was cancelled, the tool saying plainly that USBPcap is still
    not installed and cleaning up its download folder.

Not tested: a real instrument, a real recording, USBPcap's driver
actually installed (there is no USBPcap driver for Windows on ARM,
so it was not installed), and an x64 Windows PC.

What's new in 1.5.2

Changed after the second report from a user: the window closed
before the instructions could be read. Still untested with a real
recording.

  • The window stays open. Double-clicked on Windows, the window
    closed the moment the tool finished, so its help vanished and it
    looked like a crash. Every ending (normal end, nothing recorded,
    q, Ctrl-C, an error) now finishes with "Press Enter to close this
    window." It waits only in a console window of its own; from a
    command window, piped, redirected or under CI it does not.
    --pause and --no-pause force it either way.
  • WHAT TO DO NOW: when USBPcap, administrator rights, X-Rite's
    driver or the instrument is missing, a short box with the next
    steps is the last thing on the screen, with the name and folder
    of the report that holds the full instructions. When USBPcap is
    missing it offers to open Wireshark's download page (only on a
    yes).
  • Administrator rights are checked before the steps: USBPcap
    lets only administrators listen to the USB ports, so the tool now
    says so first and explains how to start it as administrator,
    instead of letting the session begin.
  • An unexpected error says sorry, saves what was collected with
    the technical details, says where, and waits.
  • i1Pro 2 on Windows 11 25H2 or newer: when X-Rite's driver is
    installed and Windows reports the instrument as working, the tool
    says so and asks to carry on and tell us whether i1Profiler
    measures with it.

What's new in 1.5.1

Changed after the first report from a user. Still untested with a
real recording.

  • Clearer help when USBPcap is missing: what it is, why it is
    needed, numbered install steps (Wireshark's "USB Capture" page,
    the restart, run the tool again), and that no security setting
    has to be switched off. When nothing could be recorded, the run
    now ends by saying so, with what to do, instead of "ALL DONE".
  • The Windows 11 25H2 driver note only where it applies: X-Rite
    says the i1Pro 2 / iO2 driver is blocked on 25H2 and newer, and
    that the i1Pro 3 works with the latest Windows 11 releases.
  • A new check for the instrument's driver: on the ArgyllCMS
    driver (as ArgyllCMS, ChromIQ or Zadig install it) the vendor
    program cannot see the instrument. The tool says how to give it
    X-Rite's driver back, and how to switch back for ChromIQ later.
  • Original i1Pro: X-Rite removed revisions A to D from
    i1Profiler in 3.2.0 (3.1.1 is needed); with both an i1Pro and an
    i1Pro 2, plug in only the one you record.
  • Devices Windows only remembers no longer count as plugged in;
    chart names containing "Studio" are no longer tagged as iO charts.

What's new in 1.5

  • A guided recording session for handheld instruments. When the
    tool sees an i1Pro, i1Pro 2, i1Pro 3, ColorMunki or i1Studio and no
    i1iO, it offers it (or start it with --i1pro). Up to nine short
    steps in the program you normally measure with (i1Profiler,
    ccStudio, the ColorMunki or i1Studio software): idle, calibrate,
    one strip, the same strip in dual scan, faster, slower, a scan
    that fails on purpose, one single patch, and without the ruler.
    Each step goes into its own file; skip, repeat or stop at any time
    without losing finished steps. After each step you note what
    happened and what the instrument's lights did.
  • What it records: the USB conversation between the program and
    the instrument, which necessarily includes the instrument's serial
    number and calibration data (the tool says so first). Only the
    instrument's own traffic goes into the folder you would send;
    nothing is uploaded, you send it yourself if you want to: in
    the forum thread where you heard about the tool, as a new
    issue, or in a private message on the forum.
  • Windows checks first (read only): Windows version, USBPcap,
    administrator rights, which USB port group the instrument is on,
    installed programs and versions. Facts, not obstacles:
    i1Profiler 3.8.7 dropped the i1Pro 2 (3.8.6 or older needed);
    X-Rite says the i1Pro 2 driver is blocked on Windows 11 25H2 and
    newer; no driver on Windows on ARM.
  • --analyze reads Windows and Linux recordings too, decodes the
    commands of the i1Pro family and the ColorMunki / i1Studio from the
    facts ArgyllCMS documents, and says in plain words whether a
    recording holds what is needed.

Nothing is promised. A recording helps us understand how the
vendor programs drive these instruments. Whether an open-source
measurement engine can ever use what we learn is a separate
question, and the answer may be no.

What has NOT been tested

  • Any recording of a real instrument. No i1Pro, i1Pro 2, i1Pro 3,
    ColorMunki or i1Studio was available while building 1.5 and 1.5.1, and no
    i1iO for 1.4. Everything about real instrument traffic is untested;
    the analyser was tested with synthetic recordings built from the
    formats ArgyllCMS documents.
  • The recording on a real Windows PC. It was tested with a
    stand-in for USBPcap's recorder; the Windows checks were tested
    against text written from the documented formats.
  • Any recording on a Mac with SIP switched off.

Owners of an i1Pro 2 or an i1iO 2: i1Profiler 3.8.7 (July 2026)
removed support for both. Stay on i1Profiler 3.8.6 or older and do
not let it update itself.

Running it on macOS or Linux

Downloaded files aren't marked as runnable, and macOS refuses
unsigned programs. The reliable way: open Terminal and run

cd ~/Downloads
chmod +x chromiq-io-probe-macos      # or chromiq-io-probe-linux
xattr -d com.apple.quarantine chromiq-io-probe-macos   # macOS only
./chromiq-io-probe-macos

chmod +x marks the file as runnable; the xattr line ...

Read more