Flow breaks when constant name changes in design_top.json #3
Assignees
Comments
|
Note that "const_186" is simply a name of the constant being using. Ideally subsequent scripts would not rely on the instance name for parsing. |
|
Yes; to be clear, this is exactly what we want to fix. Subsequent-script authors please verify that you do not rely on this name to always/only be "const"
…On April 8, 2017 4:51:23 PM PDT, Jeff Setter ***@***.***> wrote:
Note that "const_186" is simply a name of the constant being using.
Ideally subsequent scripts would not rely on the instance name for
parsing.
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
|
|
Remember Ross wrote the mapper, not me... I looked at his code, and he does hardcode a "const" in a few places, but it wasn't obvious to me that that was the problem. It could be. IMO @rdaly525 and @cdonovick should update their code to error out if the input doesn't match what is expected. As of right now, it looks like their code is silently failing, which makes this hard to track down. I would maybe also recommend removing the workaround... keeping the build working is a good idea, but IMO this is a real problem we should track down. But thanks for doing this Steven! Is everyone getting these alerts? |
|
It's possible btw that this has nothing to do with Jeff's change, remember, we pull down all the repos each time so it could have been another change. |
|
I tried the flow twice in quick succession, with the only difference
being constant name "const_186" (build 150, failed) vs. "const"
(Build 151, passed) in "design_top.json", so I'm pretty sure that
new change is causing the problem downstream. Also if you look at
past builds that succeed, all use "const" as the constant name.
If you want to do the experiment yourself, simply uncomment my
travis script hack...
# FIXME/TODO oops Jeff's recent change broke the flow.
# This temporary hack undoes the change and lets the
script succeed,
# but this is not a reasonable fix going forward!
- mv build/design_top.json /tmp/design_top.json
- sed 's/const_186/const/' /tmp/design_top.json >
build/design_top.json
- grep const build/design_top.json /tmp/design_top.json
...and you can see it fail again...
SR
PS FWIW it looks to me like the mapper script more or less passes
the constant name through unchanged, from "design_top.json" to
"mapped.json," so I suspect that it's in the PNR script where things
fall apart...
On 4/8/2017 5:31 PM, jameshegarty
wrote:
It's possible btw that this has nothing to do with Jeff's
change, remember, we pull down all the repos each time so it
could have been another change.
|
|
If changing the name of the constant changes value of
instance.get_config_value('op') from 'const' to something else then it will
break my code. However, as the API is now there is no way for me to work
around that.
Caleb
…On Apr 8, 2017 7:49 PM, "steveri" ***@***.***> wrote:
I tried the flow twice in quick succession, with the only difference
being constant name "const_186" (build 150, failed) vs. "const"
(Build 151, passed) in "design_top.json", so I'm pretty sure that
new change is causing the problem downstream. Also if you look at
past builds that succeed, all use "const" as the constant name.
If you want to do the experiment yourself, simply uncomment my
travis script hack...
# FIXME/TODO oops Jeff's recent change broke the flow.
# This temporary hack undoes the change and lets the
script succeed,
# but this is not a reasonable fix going forward!
- mv build/design_top.json /tmp/design_top.json
- sed 's/const_186/const/' /tmp/design_top.json >
build/design_top.json
- grep const build/design_top.json /tmp/design_top.json
...and you can see it fail again...
SR
PS FWIW it looks to me like the mapper script more or less passes
the constant name through unchanged, from "design_top.json" to
"mapped.json," so I suspect that it's in the PNR script where things
fall apart...
On 4/8/2017 5:31 PM, jameshegarty
wrote:
It's possible btw that this has nothing to do with Jeff's
change, remember, we pull down all the repos each time so it
could have been another change.
—
You are receiving this because you were assigned.
Reply to this email directly, view
it on GitHub, or mute
the thread.
{"api_version":"1.0","publisher":{"api_key":"
05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":
{"external_key":"github/StanfordAHA/CGRAFlow","title":
"StanfordAHA/CGRAFlow","subtitle":"GitHub repository","main_image_url":"
https://cloud.githubusercontent.com/assets/143418/17495839/a5054eac-5d88-
11e6-95fc-7290892c7bb5.png","avatar_image_url":"https://
cloud.githubusercontent.com/assets/143418/15842166/
7c72db34-2c0b-11e6-9aed-b52498112777.png","action":{"name":"Open in
GitHub","url":"https://github.com/StanfordAHA/CGRAFlow"}},"
***@***.*** in #3:
It's possible btw that this has nothing to do with Jeff's change, remember,
we pull down all the repos each time so it could have been another
change."}],"action":{"name":"View Issue","url":"https://github.
com/StanfordAHA/CGRAFlow/issues/3#issuecomment-292755048"}}}
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AF13dCCYgaJ7O0cIgi3gVqf8QZYIetbPks5ruEclgaJpZM4M33iU>
.
|
|
It appears to be producing this (see build log 149 https://travis-ci.org/StanfordAHA/CGRAFlow/builds/220134733):
instances": {
"const_186": {
"config": {
"constvalue": 2,
"op": "const"
…On April 8, 2017 9:10:43 PM PDT, Caleb Donovick ***@***.***> wrote:
If changing the name of the constant changes value of
instance.get_config_value('op') from 'const' to something else then it
will
break my code. However, as the API is now there is no way for me to
work
around that.
Caleb
On Apr 8, 2017 7:49 PM, "steveri" ***@***.***> wrote:
>
>
>
>
>
> I tried the flow twice in quick succession, with the only difference
> being constant name "const_186" (build 150, failed) vs. "const"
> (Build 151, passed) in "design_top.json", so I'm pretty sure that
> new change is causing the problem downstream. Also if you look at
> past builds that succeed, all use "const" as the constant name.
>
> If you want to do the experiment yourself, simply uncomment my
> travis script hack...
>
> # FIXME/TODO oops Jeff's recent change broke the flow.
> # This temporary hack undoes the change and lets the
> script succeed,
> # but this is not a reasonable fix going forward!
> - mv build/design_top.json /tmp/design_top.json
> - sed 's/const_186/const/' /tmp/design_top.json >
> build/design_top.json
> - grep const build/design_top.json /tmp/design_top.json
>
> ...and you can see it fail again...
>
> SR
>
> PS FWIW it looks to me like the mapper script more or less passes
> the constant name through unchanged, from "design_top.json" to
> "mapped.json," so I suspect that it's in the PNR script where things
> fall apart...
>
>
> On 4/8/2017 5:31 PM, jameshegarty
> wrote:
>
>
> It's possible btw that this has nothing to do with Jeff's
> change, remember, we pull down all the repos each time so it
> could have been another change.
> —
> You are receiving this because you were assigned.
> Reply to this email directly, view
> it on GitHub, or mute
> the thread.
>
>
>
>
>
>
>
> {"api_version":"1.0","publisher":{"api_key":"
> 05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":
> {"external_key":"github/StanfordAHA/CGRAFlow","title":
> "StanfordAHA/CGRAFlow","subtitle":"GitHub
repository","main_image_url":"
>
https://cloud.githubusercontent.com/assets/143418/17495839/a5054eac-5d88-
> 11e6-95fc-7290892c7bb5.png","avatar_image_url":"https://
> cloud.githubusercontent.com/assets/143418/15842166/
> 7c72db34-2c0b-11e6-9aed-b52498112777.png","action":{"name":"Open in
> GitHub","url":"https://github.com/StanfordAHA/CGRAFlow"}},"
> ***@***.*** in
#3:
> It's possible btw that this has nothing to do with Jeff's change,
remember,
> we pull down all the repos each time so it could have been another
> change."}],"action":{"name":"View Issue","url":"https://github.
> com/StanfordAHA/CGRAFlow/issues/3#issuecomment-292755048"}}}
>
> —
> You are receiving this because you were mentioned.
> Reply to this email directly, view it on GitHub
>
<#3 (comment)>,
> or mute the thread
>
<https://github.com/notifications/unsubscribe-auth/AF13dCCYgaJ7O0cIgi3gVqf8QZYIetbPks5ruEclgaJpZM4M33iU>
> .
>
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
|
|
Oh and also this, earlier in the same file
"instances": {
"const_186": {
"config": {
"value": 2
},
…On April 8, 2017 9:41:47 PM PDT, Stephen ***@***.***> wrote:
It appears to be producing this (see build log 149
https://travis-ci.org/StanfordAHA/CGRAFlow/builds/220134733):
instances": {
"const_186": {
"config": {
"constvalue": 2,
"op": "const"
On April 8, 2017 9:10:43 PM PDT, Caleb Donovick
***@***.***> wrote:
>If changing the name of the constant changes value of
>instance.get_config_value('op') from 'const' to something else then it
>will
>break my code. However, as the API is now there is no way for me to
>work
>around that.
>
>
>Caleb
>
>On Apr 8, 2017 7:49 PM, "steveri" ***@***.***> wrote:
>
>>
>>
>>
>>
>>
>> I tried the flow twice in quick succession, with the only difference
>> being constant name "const_186" (build 150, failed) vs. "const"
>> (Build 151, passed) in "design_top.json", so I'm pretty sure that
>> new change is causing the problem downstream. Also if you look at
>> past builds that succeed, all use "const" as the constant name.
>>
>> If you want to do the experiment yourself, simply uncomment my
>> travis script hack...
>>
>> # FIXME/TODO oops Jeff's recent change broke the flow.
>> # This temporary hack undoes the change and lets the
>> script succeed,
>> # but this is not a reasonable fix going forward!
>> - mv build/design_top.json /tmp/design_top.json
>> - sed 's/const_186/const/' /tmp/design_top.json >
>> build/design_top.json
>> - grep const build/design_top.json /tmp/design_top.json
>>
>> ...and you can see it fail again...
>>
>> SR
>>
>> PS FWIW it looks to me like the mapper script more or less passes
>> the constant name through unchanged, from "design_top.json" to
>> "mapped.json," so I suspect that it's in the PNR script where things
>> fall apart...
>>
>>
>> On 4/8/2017 5:31 PM, jameshegarty
>> wrote:
>>
>>
>> It's possible btw that this has nothing to do with Jeff's
>> change, remember, we pull down all the repos each time so it
>> could have been another change.
>> —
>> You are receiving this because you were assigned.
>> Reply to this email directly, view
>> it on GitHub, or mute
>> the thread.
>>
>>
>>
>>
>>
>>
>>
>> {"api_version":"1.0","publisher":{"api_key":"
>> 05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":
>> {"external_key":"github/StanfordAHA/CGRAFlow","title":
>> "StanfordAHA/CGRAFlow","subtitle":"GitHub
>repository","main_image_url":"
>>
>https://cloud.githubusercontent.com/assets/143418/17495839/a5054eac-5d88-
>> 11e6-95fc-7290892c7bb5.png","avatar_image_url":"https://
>> cloud.githubusercontent.com/assets/143418/15842166/
>> 7c72db34-2c0b-11e6-9aed-b52498112777.png","action":{"name":"Open in
>> GitHub","url":"https://github.com/StanfordAHA/CGRAFlow"}},"
>> ***@***.*** in
>#3:
>> It's possible btw that this has nothing to do with Jeff's change,
>remember,
>> we pull down all the repos each time so it could have been another
>> change."}],"action":{"name":"View Issue","url":"https://github.
>> com/StanfordAHA/CGRAFlow/issues/3#issuecomment-292755048"}}}
>>
>> —
>> You are receiving this because you were mentioned.
>> Reply to this email directly, view it on GitHub
>>
><#3 (comment)>,
>> or mute the thread
>>
><https://github.com/notifications/unsubscribe-auth/AF13dCCYgaJ7O0cIgi3gVqf8QZYIetbPks5ruEclgaJpZM4M33iU>
>> .
>>
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
|
|
The latter of the two ?might? break my code but I'm not completely up on
COREIR format. Looked 150 like it broke because missed named filles.
Caleb
…On Apr 8, 2017 9:47 PM, "steveri" ***@***.***> wrote:
Oh and also this, earlier in the same file
"instances": {
"const_186": {
"config": {
"value": 2
},
On April 8, 2017 9:41:47 PM PDT, Stephen ***@***.***> wrote:
>It appears to be producing this (see build log 149
>https://travis-ci.org/StanfordAHA/CGRAFlow/builds/220134733):
>
>
>instances": {
> "const_186": {
> "config": {
> "constvalue": 2,
> "op": "const"
>
>
>
>On April 8, 2017 9:10:43 PM PDT, Caleb Donovick
***@***.***> wrote:
>>If changing the name of the constant changes value of
>>instance.get_config_value('op') from 'const' to something else then it
>>will
>>break my code. However, as the API is now there is no way for me to
>>work
>>around that.
>>
>>
>>Caleb
>>
>>On Apr 8, 2017 7:49 PM, "steveri" ***@***.***> wrote:
>>
>>>
>>>
>>>
>>>
>>>
>>> I tried the flow twice in quick succession, with the only difference
>>> being constant name "const_186" (build 150, failed) vs. "const"
>>> (Build 151, passed) in "design_top.json", so I'm pretty sure that
>>> new change is causing the problem downstream. Also if you look at
>>> past builds that succeed, all use "const" as the constant name.
>>>
>>> If you want to do the experiment yourself, simply uncomment my
>>> travis script hack...
>>>
>>> # FIXME/TODO oops Jeff's recent change broke the flow.
>>> # This temporary hack undoes the change and lets the
>>> script succeed,
>>> # but this is not a reasonable fix going forward!
>>> - mv build/design_top.json /tmp/design_top.json
>>> - sed 's/const_186/const/' /tmp/design_top.json >
>>> build/design_top.json
>>> - grep const build/design_top.json /tmp/design_top.json
>>>
>>> ...and you can see it fail again...
>>>
>>> SR
>>>
>>> PS FWIW it looks to me like the mapper script more or less passes
>>> the constant name through unchanged, from "design_top.json" to
>>> "mapped.json," so I suspect that it's in the PNR script where things
>>> fall apart...
>>>
>>>
>>> On 4/8/2017 5:31 PM, jameshegarty
>>> wrote:
>>>
>>>
>>> It's possible btw that this has nothing to do with Jeff's
>>> change, remember, we pull down all the repos each time so it
>>> could have been another change.
>>> —
>>> You are receiving this because you were assigned.
>>> Reply to this email directly, view
>>> it on GitHub, or mute
>>> the thread.
>>>
>>>
>>>
>>>
>>>
>>>
>>>
>>> {"api_version":"1.0","publisher":{"api_key":"
>>> 05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":
>>> {"external_key":"github/StanfordAHA/CGRAFlow","title":
>>> "StanfordAHA/CGRAFlow","subtitle":"GitHub
>>repository","main_image_url":"
>>>
>>https://cloud.githubusercontent.com/assets/
143418/17495839/a5054eac-5d88-
>>> 11e6-95fc-7290892c7bb5.png","avatar_image_url":"https://
>>> cloud.githubusercontent.com/assets/143418/15842166/
>>> 7c72db34-2c0b-11e6-9aed-b52498112777.png","action":{"name":"Open in
>>> GitHub","url":"https://github.com/StanfordAHA/CGRAFlow"}},"
>>> ***@***.*** in
>>#3:
>>> It's possible btw that this has nothing to do with Jeff's change,
>>remember,
>>> we pull down all the repos each time so it could have been another
>>> change."}],"action":{"name":"View Issue","url":"https://github.
>>> com/StanfordAHA/CGRAFlow/issues/3#issuecomment-292755048"}}}
>>>
>>> —
>>> You are receiving this because you were mentioned.
>>> Reply to this email directly, view it on GitHub
>>>
>><#3 (comment)
>,
>>> or mute the thread
>>>
>><https://github.com/notifications/unsubscribe-auth/
AF13dCCYgaJ7O0cIgi3gVqf8QZYIetbPks5ruEclgaJpZM4M33iU>
>>> .
>>>
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#3 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AF13dFqW-N6Tc6EkA_WEHjpajPWCEXioks5ruGLDgaJpZM4M33iU>
.
|
|
I'll look at all of it tomorrow, trying to read Travis logs on a phone is
suboptimal.
…On Apr 8, 2017 9:49 PM, "Caleb Donovick" ***@***.***> wrote:
The latter of the two ?might? break my code but I'm not completely up on
COREIR format. Looked 150 like it broke because missed named filles.
Caleb
On Apr 8, 2017 9:47 PM, "steveri" ***@***.***> wrote:
> Oh and also this, earlier in the same file
>
> "instances": {
> "const_186": {
> "config": {
> "value": 2
> },
>
>
> On April 8, 2017 9:41:47 PM PDT, Stephen ***@***.***> wrote:
> >It appears to be producing this (see build log 149
> >https://travis-ci.org/StanfordAHA/CGRAFlow/builds/220134733):
> >
> >
> >instances": {
> > "const_186": {
> > "config": {
> > "constvalue": 2,
> > "op": "const"
> >
> >
> >
> >On April 8, 2017 9:10:43 PM PDT, Caleb Donovick
> ***@***.***> wrote:
> >>If changing the name of the constant changes value of
> >>instance.get_config_value('op') from 'const' to something else then it
> >>will
> >>break my code. However, as the API is now there is no way for me to
> >>work
> >>around that.
> >>
> >>
> >>Caleb
> >>
> >>On Apr 8, 2017 7:49 PM, "steveri" ***@***.***> wrote:
> >>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> I tried the flow twice in quick succession, with the only difference
> >>> being constant name "const_186" (build 150, failed) vs. "const"
> >>> (Build 151, passed) in "design_top.json", so I'm pretty sure that
> >>> new change is causing the problem downstream. Also if you look at
> >>> past builds that succeed, all use "const" as the constant name.
> >>>
> >>> If you want to do the experiment yourself, simply uncomment my
> >>> travis script hack...
> >>>
> >>> # FIXME/TODO oops Jeff's recent change broke the flow.
> >>> # This temporary hack undoes the change and lets the
> >>> script succeed,
> >>> # but this is not a reasonable fix going forward!
> >>> - mv build/design_top.json /tmp/design_top.json
> >>> - sed 's/const_186/const/' /tmp/design_top.json >
> >>> build/design_top.json
> >>> - grep const build/design_top.json /tmp/design_top.json
> >>>
> >>> ...and you can see it fail again...
> >>>
> >>> SR
> >>>
> >>> PS FWIW it looks to me like the mapper script more or less passes
> >>> the constant name through unchanged, from "design_top.json" to
> >>> "mapped.json," so I suspect that it's in the PNR script where things
> >>> fall apart...
> >>>
> >>>
> >>> On 4/8/2017 5:31 PM, jameshegarty
> >>> wrote:
> >>>
> >>>
> >>> It's possible btw that this has nothing to do with Jeff's
> >>> change, remember, we pull down all the repos each time so it
> >>> could have been another change.
> >>> —
> >>> You are receiving this because you were assigned.
> >>> Reply to this email directly, view
> >>> it on GitHub, or mute
> >>> the thread.
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> {"api_version":"1.0","publisher":{"api_key":"
> >>> 05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":
> >>> {"external_key":"github/StanfordAHA/CGRAFlow","title":
> >>> "StanfordAHA/CGRAFlow","subtitle":"GitHub
> >>repository","main_image_url":"
> >>>
> >>https://cloud.githubusercontent.com/assets/143418/
> 17495839/a5054eac-5d88-
> >>> 11e6-95fc-7290892c7bb5.png","avatar_image_url":"https://
> >>> cloud.githubusercontent.com/assets/143418/15842166/
> >>> 7c72db34-2c0b-11e6-9aed-b52498112777.png","action":{"name":"Open in
> >>> GitHub","url":"https://github.com/StanfordAHA/CGRAFlow"}},"
> >>> ***@***.*** in
> >>#3:
> >>> It's possible btw that this has nothing to do with Jeff's change,
> >>remember,
> >>> we pull down all the repos each time so it could have been another
> >>> change."}],"action":{"name":"View Issue","url":"https://github.
> >>> com/StanfordAHA/CGRAFlow/issues/3#issuecomment-292755048"}}}
> >>>
> >>> —
> >>> You are receiving this because you were mentioned.
> >>> Reply to this email directly, view it on GitHub
> >>>
> >><#3 (comment)
> ment-292759941>,
> >>> or mute the thread
> >>>
> >><https://github.com/notifications/unsubscribe-auth/AF13dCC
> YgaJ7O0cIgi3gVqf8QZYIetbPks5ruEclgaJpZM4M33iU>
> >>> .
> >>>
>
> --
> Sent from my Android device with K-9 Mail. Please excuse my brevity.
>
> —
> You are receiving this because you were mentioned.
> Reply to this email directly, view it on GitHub
> <#3 (comment)>,
> or mute the thread
> <https://github.com/notifications/unsubscribe-auth/AF13dFqW-N6Tc6EkA_WEHjpajPWCEXioks5ruGLDgaJpZM4M33iU>
> .
>
|
|
Yes dont look at build 150 look at 149 instead.
…On April 8, 2017 9:50:00 PM PDT, Caleb Donovick ***@***.***> wrote:
The latter of the two ?might? break my code but I'm not completely up
on
COREIR format. Looked 150 like it broke because missed named filles.
Caleb
On Apr 8, 2017 9:47 PM, "steveri" ***@***.***> wrote:
> Oh and also this, earlier in the same file
>
> "instances": {
> "const_186": {
> "config": {
> "value": 2
> },
>
>
> On April 8, 2017 9:41:47 PM PDT, Stephen ***@***.***> wrote:
> >It appears to be producing this (see build log 149
> >https://travis-ci.org/StanfordAHA/CGRAFlow/builds/220134733):
> >
> >
> >instances": {
> > "const_186": {
> > "config": {
> > "constvalue": 2,
> > "op": "const"
> >
> >
> >
> >On April 8, 2017 9:10:43 PM PDT, Caleb Donovick
> ***@***.***> wrote:
> >>If changing the name of the constant changes value of
> >>instance.get_config_value('op') from 'const' to something else then
it
> >>will
> >>break my code. However, as the API is now there is no way for me to
> >>work
> >>around that.
> >>
> >>
> >>Caleb
> >>
> >>On Apr 8, 2017 7:49 PM, "steveri" ***@***.***> wrote:
> >>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> I tried the flow twice in quick succession, with the only
difference
> >>> being constant name "const_186" (build 150, failed) vs. "const"
> >>> (Build 151, passed) in "design_top.json", so I'm pretty sure that
> >>> new change is causing the problem downstream. Also if you look at
> >>> past builds that succeed, all use "const" as the constant name.
> >>>
> >>> If you want to do the experiment yourself, simply uncomment my
> >>> travis script hack...
> >>>
> >>> # FIXME/TODO oops Jeff's recent change broke the flow.
> >>> # This temporary hack undoes the change and lets the
> >>> script succeed,
> >>> # but this is not a reasonable fix going forward!
> >>> - mv build/design_top.json /tmp/design_top.json
> >>> - sed 's/const_186/const/' /tmp/design_top.json >
> >>> build/design_top.json
> >>> - grep const build/design_top.json /tmp/design_top.json
> >>>
> >>> ...and you can see it fail again...
> >>>
> >>> SR
> >>>
> >>> PS FWIW it looks to me like the mapper script more or less passes
> >>> the constant name through unchanged, from "design_top.json" to
> >>> "mapped.json," so I suspect that it's in the PNR script where
things
> >>> fall apart...
> >>>
> >>>
> >>> On 4/8/2017 5:31 PM, jameshegarty
> >>> wrote:
> >>>
> >>>
> >>> It's possible btw that this has nothing to do with Jeff's
> >>> change, remember, we pull down all the repos each time so it
> >>> could have been another change.
> >>> —
> >>> You are receiving this because you were assigned.
> >>> Reply to this email directly, view
> >>> it on GitHub, or mute
> >>> the thread.
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> {"api_version":"1.0","publisher":{"api_key":"
> >>> 05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":
> >>> {"external_key":"github/StanfordAHA/CGRAFlow","title":
> >>> "StanfordAHA/CGRAFlow","subtitle":"GitHub
> >>repository","main_image_url":"
> >>>
> >>https://cloud.githubusercontent.com/assets/
> 143418/17495839/a5054eac-5d88-
> >>> 11e6-95fc-7290892c7bb5.png","avatar_image_url":"https://
> >>> cloud.githubusercontent.com/assets/143418/15842166/
> >>> 7c72db34-2c0b-11e6-9aed-b52498112777.png","action":{"name":"Open
in
> >>> GitHub","url":"https://github.com/StanfordAHA/CGRAFlow"}},"
> >>> ***@***.***
in
> >>#3:
> >>> It's possible btw that this has nothing to do with Jeff's change,
> >>remember,
> >>> we pull down all the repos each time so it could have been
another
> >>> change."}],"action":{"name":"View Issue","url":"https://github.
> >>> com/StanfordAHA/CGRAFlow/issues/3#issuecomment-292755048"}}}
> >>>
> >>> —
> >>> You are receiving this because you were mentioned.
> >>> Reply to this email directly, view it on GitHub
> >>>
>
>><#3 (comment)
> >,
> >>> or mute the thread
> >>>
> >><https://github.com/notifications/unsubscribe-auth/
> AF13dCCYgaJ7O0cIgi3gVqf8QZYIetbPks5ruEclgaJpZM4M33iU>
> >>> .
> >>>
>
> --
> Sent from my Android device with K-9 Mail. Please excuse my brevity.
>
> —
> You are receiving this because you were mentioned.
> Reply to this email directly, view it on GitHub
>
<#3 (comment)>,
> or mute the thread
>
<https://github.com/notifications/unsubscribe-auth/AF13dFqW-N6Tc6EkA_WEHjpajPWCEXioks5ruGLDgaJpZM4M33iU>
> .
>
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
|
|
Agreed, me too, my eyes are starting to tear up. :)
Certainly doesnt have to be fixed tonight.
SR
…On April 8, 2017 9:51:37 PM PDT, Caleb Donovick ***@***.***> wrote:
I'll look at all of it tomorrow, trying to read Travis logs on a phone
is
suboptimal.
On Apr 8, 2017 9:49 PM, "Caleb Donovick" ***@***.***> wrote:
> The latter of the two ?might? break my code but I'm not completely up
on
> COREIR format. Looked 150 like it broke because missed named filles.
>
>
> Caleb
>
> On Apr 8, 2017 9:47 PM, "steveri" ***@***.***> wrote:
>
>> Oh and also this, earlier in the same file
>>
>> "instances": {
>> "const_186": {
>> "config": {
>> "value": 2
>> },
>>
>>
>> On April 8, 2017 9:41:47 PM PDT, Stephen ***@***.***>
wrote:
>> >It appears to be producing this (see build log 149
>> >https://travis-ci.org/StanfordAHA/CGRAFlow/builds/220134733):
>> >
>> >
>> >instances": {
>> > "const_186": {
>> > "config": {
>> > "constvalue": 2,
>> > "op": "const"
>> >
>> >
>> >
>> >On April 8, 2017 9:10:43 PM PDT, Caleb Donovick
>> ***@***.***> wrote:
>> >>If changing the name of the constant changes value of
>> >>instance.get_config_value('op') from 'const' to something else
then it
>> >>will
>> >>break my code. However, as the API is now there is no way for me
to
>> >>work
>> >>around that.
>> >>
>> >>
>> >>Caleb
>> >>
>> >>On Apr 8, 2017 7:49 PM, "steveri" ***@***.***>
wrote:
>> >>
>> >>>
>> >>>
>> >>>
>> >>>
>> >>>
>> >>> I tried the flow twice in quick succession, with the only
difference
>> >>> being constant name "const_186" (build 150, failed) vs. "const"
>> >>> (Build 151, passed) in "design_top.json", so I'm pretty sure
that
>> >>> new change is causing the problem downstream. Also if you look
at
>> >>> past builds that succeed, all use "const" as the constant name.
>> >>>
>> >>> If you want to do the experiment yourself, simply uncomment my
>> >>> travis script hack...
>> >>>
>> >>> # FIXME/TODO oops Jeff's recent change broke the flow.
>> >>> # This temporary hack undoes the change and lets the
>> >>> script succeed,
>> >>> # but this is not a reasonable fix going forward!
>> >>> - mv build/design_top.json /tmp/design_top.json
>> >>> - sed 's/const_186/const/' /tmp/design_top.json >
>> >>> build/design_top.json
>> >>> - grep const build/design_top.json /tmp/design_top.json
>> >>>
>> >>> ...and you can see it fail again...
>> >>>
>> >>> SR
>> >>>
>> >>> PS FWIW it looks to me like the mapper script more or less
passes
>> >>> the constant name through unchanged, from "design_top.json" to
>> >>> "mapped.json," so I suspect that it's in the PNR script where
things
>> >>> fall apart...
>> >>>
>> >>>
>> >>> On 4/8/2017 5:31 PM, jameshegarty
>> >>> wrote:
>> >>>
>> >>>
>> >>> It's possible btw that this has nothing to do with Jeff's
>> >>> change, remember, we pull down all the repos each time so it
>> >>> could have been another change.
>> >>> —
>> >>> You are receiving this because you were assigned.
>> >>> Reply to this email directly, view
>> >>> it on GitHub, or mute
>> >>> the thread.
>> >>>
>> >>>
>> >>>
>> >>>
>> >>>
>> >>>
>> >>>
>> >>> {"api_version":"1.0","publisher":{"api_key":"
>> >>> 05dde50f1d1a384dd78767c55493e4bb","name":"GitHub"},"entity":
>> >>> {"external_key":"github/StanfordAHA/CGRAFlow","title":
>> >>> "StanfordAHA/CGRAFlow","subtitle":"GitHub
>> >>repository","main_image_url":"
>> >>>
>> >>https://cloud.githubusercontent.com/assets/143418/
>> 17495839/a5054eac-5d88-
>> >>> 11e6-95fc-7290892c7bb5.png","avatar_image_url":"https://
>> >>> cloud.githubusercontent.com/assets/143418/15842166/
>> >>> 7c72db34-2c0b-11e6-9aed-b52498112777.png","action":{"name":"Open
in
>> >>> GitHub","url":"https://github.com/StanfordAHA/CGRAFlow"}},"
>> >>> ***@***.***
in
>> >>#3:
>> >>> It's possible btw that this has nothing to do with Jeff's
change,
>> >>remember,
>> >>> we pull down all the repos each time so it could have been
another
>> >>> change."}],"action":{"name":"View Issue","url":"https://github.
>> >>> com/StanfordAHA/CGRAFlow/issues/3#issuecomment-292755048"}}}
>> >>>
>> >>> —
>> >>> You are receiving this because you were mentioned.
>> >>> Reply to this email directly, view it on GitHub
>> >>>
>> >><#3 (comment)
>> ment-292759941>,
>> >>> or mute the thread
>> >>>
>> >><https://github.com/notifications/unsubscribe-auth/AF13dCC
>> YgaJ7O0cIgi3gVqf8QZYIetbPks5ruEclgaJpZM4M33iU>
>> >>> .
>> >>>
>>
>> --
>> Sent from my Android device with K-9 Mail. Please excuse my brevity.
>>
>> —
>> You are receiving this because you were mentioned.
>> Reply to this email directly, view it on GitHub
>>
<#3 (comment)>,
>> or mute the thread
>>
<https://github.com/notifications/unsubscribe-auth/AF13dFqW-N6Tc6EkA_WEHjpajPWCEXioks5ruGLDgaJpZM4M33iU>
>> .
>>
>
--
Sent from my Android device with K-9 Mail. Please excuse my brevity.
|
|
FYI I removed the log for build 150 so as to (hopefully) lessen
confusion (it broke because I initially messed up my sed-script
hack). For an example of the failing design_top and mapped.json
files, see log 149 instead...
|
|
For completeness, note, this (below) should have referred to "build
149, failed" and not "build 150, failed." Build 150 is not
useful and I deleted it as per previous e-mail...
SR
On 4/8/2017 7:49 PM, Stephen Richardson
wrote:
I tried the flow twice in quick succession, with the only difference being constant name "const_186" (build 150, failed) vs. "const" (Build 151, passed) in "design_top.json", so I'm pretty sure that new change is causing the problem downstream. Also if you look at past builds that succeed, all use "const" as the constant name.
If you want to do the experiment yourself, simply uncomment my
travis script hack...
# FIXME/TODO oops Jeff's recent change broke the flow.
# This temporary hack undoes the change and lets the
script succeed,
# but this is not a reasonable fix going forward!
- mv build/design_top.json /tmp/design_top.json
- sed 's/const_186/const/' /tmp/design_top.json >
build/design_top.json
- grep const build/design_top.json /tmp/design_top.json
...and you can see it fail again...
SR
PS FWIW it looks to me like the mapper script more or less passes
the constant name through unchanged, from "design_top.json" to
"mapped.json," so I suspect that it's in the PNR script where
things fall apart...
|
|
As recommended earlier by James, I am going to remove my sed hack from the travis script so that it will fail until we get this fixed...
|
|
BTW if (like me) you are annoyed that everyone can see your comments except you, note that there is a place in your account settings where you can set "include your own updates" (seems to be off by default) - see https://github.com/settings/notifications |
|
I am going to push a change that should fix this problem, but break Jeff's code. |
|
I pushed the necessary change to my repo. It appears that it is working again. |
|
Great! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
When Jeff changes the name of the const from "const" to "const_186" the flow breaks. I (steveri) added a hack in the travis script to undo the name, but this is of course not the right fix. I suspect the problem is either in the mapper or in one of the pnr scripts, respectively. You can see source for working and non-working design_top.json in builds 145 (https://travis-ci.org/StanfordAHA/CGRAFlow/builds/219574814) and 146 (https://travis-ci.org/StanfordAHA/CGRAFlow/builds/219898992) respectively---the only difference is one line where the working version says "const" and the nonworking version says "const_186"
diff design_top_145.json design_top_146.json
< "const"
The text was updated successfully, but these errors were encountered: