[Clarification request/Improvement] Add version and github info always, also when --enable-debug was missing

[Wikinaut]

I know, that tesseract --version outputs the exact version and github commit hash, if it was compiled with ./configure --enable-debug, example:

tesseract 4.00.00alpha-278-gc768b58
 leptonica-1.74.1
  libjpeg 6b (libjpeg-turbo 1.3.1) : libpng 1.2.50 : libtiff 4.0.3 : zlib 1.2.8

 Found AVX
 Found SSE

I wonder, why this is not output, when the configuration setting was not given and I suggest to modify the build code so that the version and github commit hash is always shown with the --versioncommand.

[Shreeshrii]

Yes. I agree, adding version info with github commit hash in non-debug builds will be very helpful.

An earlier request - #581 (comment)

[Shreeshrii]

Needs change in

tesseract/api/baseapi.cpp

Line 140 in fd0683f

#if defined(GIT_REV) && (defined(DEBUG) || defined(_DEBUG))

[Wikinaut]

@zdenop Is

tesseract/api/baseapi.cpp

Line 140 in fd0683f

#if defined(GIT_REV) && (defined(DEBUG) || defined(_DEBUG))

the only change needed, and would you accept a corresponding pull request ?

[zdenop]

I am against to show GIT_REV in release mode:

• Experienced users/testers should know what they are doing, what their are testing
• Inexperienced users should use released stable version

[Wikinaut]

@zdenop Where exactly is your problem to print the GIT_REV when users use --version ?

I *have a problem when, as an expert, I run always the latest version, but sometimes after git pull the latest build is not the "installed" build. As a user I want to have the fastest code, i.e. built without --enable-debug.

I again request to think over it, and to always - if available - burn the GIT_REV into the code. I ask you for this.

[Wikinaut]

Not using --enable-debug increases processing speed about 6%.

[stweil]

@zdenop, what about using GIT_REV only for versions without tag, no matter whether it is a debug or a release build? Then released stable versions would not show GIT_REV.

[amitdo]

Not using --enable-debug increases processing speed about 6%.

Only ~6%? With LSTM mode?

#40 (comment)

[amitdo]

BTW, maybe we should add another debug/optimization option with -Og?
https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html

[Shreeshrii]

This has now been implemented by recent commits by @stweil .

tesseract -v
tesseract 4.0.0-beta.1-59-g2cc4
 leptonica-1.76.0
  libjpeg 8d (libjpeg-turbo 1.3.0) : libpng 1.2.50 : libtiff 4.0.3 : zlib 1.2.8

 Found AVX
 Found SSE

Please check with latest code from master and close the issue.

[Wikinaut]

I tried to build the new master version but make stopped with

libtool:   error: REVISION '00' must be a nonnegative integer
libtool:   error: '4:00' is not valid version information

relevant lines are:

libtool: link: ( cd ".libs" && rm -f "libtesseract_api.la" && ln -s "../libtesseract_api.la" "libtesseract_api.la" )
/bin/bash ../libtool  --tag=CXX   --mode=link g++  -g -O2 -std=c++11 -L/usr/local/lib -llept  -version-info 4:00 -no-undefined   -o libtesseract.la -rpath /usr/local/lib  libtesseract_api.la ../ccmain/libtesseract_main.la ../textord/ltesseract_textord.la ../wordrec/libtesseract_wordrec.la ../classify/libtesseract_classify.la ../dict/libtesseract_dict.la ../arch/libtesseract_arch.la ../arch/libtesseract_avx.la ../arch/libtesseract_avx2.la ../arch/libtesseract_sse.l../lstm/libtesseract_lstm.la ../ccstruct/libtesseract_ccstruct.la ../cutil/libtesseract_cutil.la ../viewer/libtesseract_viewer.la ../ccutil/libtesseract_ccutil.la ../opencl/libtesseract_opencl.la  -lpthread 
libtool:   error: REVISION '00' must be a nonnegative integer
libtool:   error: '4:00' is not valid version information
Makefile:524: die Regel für Ziel „libtesseract.la“ scheiterte
make[2]: *** [libtesseract.la] Fehler 1
make[2]: Verzeichnis „/work/usr/local/src/tesseract/api“ wird verlassen
Makefile:494: die Regel für Ziel „all-recursive“ scheiterte
make[1]: *** [all-recursive] Fehler 1
make[1]: Verzeichnis „/work/usr/local/src/tesseract“ wird verlassen
Makefile:403: die Regel für Ziel „all“ scheiterte
make: *** [all] Fehler 2

[Shreeshrii]

How are you trying to build? I used the following today, without any problem.

#!/bin/bash
./autogen.sh
./configure --disable-openmp --disable-graphics --disable-opencl
make
sudo make install
sudo ldconfig
make training
sudo make training-install

[Wikinaut]

make distclean
git pull upstream master
./autogen.sh
./configure
make

[Shreeshrii]

@stweil Please see make error above.

[stweil]

@Wikinaut, it looks like you are using git, but your latest tag does not match the expectations (MAJOR.MINOR.PATCH with MAJOR, MINOR and PATCH being numerical values, maybe followed by additional non-numeric characters). The latest tag should be 4.0.0-beta.1. What do you get if you run git describe --abbrev=4?

[Wikinaut]

I just built the master branch, as usual.

git describe --abbrev=4
4.00.00dev-787-g2cc4

[Wikinaut]

@stweil have a look (I just updated this above) to the make output

libtool: link: ( cd ".libs" && rm -f "libtesseract_api.la" && ln -s "../libtesseract_api.la" "libtesseract_api.la" )
/bin/bash ../libtool --tag=CXX --mode=link g++ -g -O2 -std=c++11 -L/usr/local/lib -llept -version-info 4:00 -no-undefined -o libtesseract.la -rpath /usr/local/lib libtesseract_api.la ../ccmain/libtesseract_main.la ../textord/ltesseract_textord.la ../wordrec/libtesseract_wordrec.la ../classify/libtesseract_classify.la ../dict/libtesseract_dict.la ../arch/libtesseract_arch.la ../arch/libtesseract_avx.la ../arch/libtesseract_avx2.la ../arch/libtesseract_sse.l../lstm/libtesseract_lstm.la ../ccstruct/libtesseract_ccstruct.la ../cutil/libtesseract_cutil.la ../viewer/libtesseract_viewer.la ../ccutil/libtesseract_ccutil.la ../opencl/libtesseract_opencl.la -lpthread

Especially to -version-info 4:00. Is the colon correct?

[stweil]

@Wikinaut, your local git has the latest code, but not the latest tags (see list of all tags). It still uses the old tag 4.00.00dev. Are you pulling from your personal fork (where the tag might be missing)? If you pull from tesseract-ocr, you should automatically get the latest tag, unless you have disabled that in your git settings. Try git pull --tags to force pulling all tags.

[stweil]

The error message which you see comes from -version-info 4:00. It should be -version-info 4:0.

[Shreeshrii]

What info do u see for git remote -v?

[Wikinaut]

@Shreeshrii

origin  git@github.com:Wikinaut/tesseract.git (fetch)
origin  git@github.com:Wikinaut/tesseract.git (push)
upstream        https://github.com/tesseract-ocr/tesseract.git (fetch)
upstream        https://github.com/tesseract-ocr/tesseract.git (push)

[Shreeshrii]

Try git pull --tags to force pulling all tags.

and then

run git describe --abbrev=4

[stweil]

git fetch --all --tags, then we are sure that you won't miss a tag. :-)

[Wikinaut]

ok, works now.

Result as wanted:

tesseract -v
tesseract 4.0.0-beta.1-59-g2cc4
 leptonica-1.76.0
  libjpeg 6b (libjpeg-turbo 1.5.1) : libpng 1.2.50 : libtiff 4.0.8 : zlib 1.2.8
 Found AVX2
 Found AVX
 Found SSE

Closing this as solved, thanks.

[amitdo]

tesseract 4.0.0-beta.1-59-g2cc4

@stweil, according to the semver 2.0.0 spec, there should be only one -. The rest should be replaced by dots.

BTW, from where the field before the commit hash comes from?
(59 here)

[Shreeshrii]

I think that is number of commits since last tag, beta.1

…

On Thu 29 Mar, 2018, 12:04 AM Amit D., ***@***.***> wrote: tesseract 4.0.0-beta.1-59-g2cc4 @stweil <https://github.com/stweil>, according to the semver 2.0.0 spec, there should be only one -. The rest should be replaced by dots. BTW, from where the field before the commit hash comes from? (59 here) — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#723 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AE2_o0bZI9uMx1qCMxYsmdDjOPypNc7sks5ti9fDgaJpZM4MDsWM> .

[stweil]

4.0.0-beta.1-59-g2cc4 is not a semver tag, so I don't think that the semver spec applies here. It would be possible to replace '-' by '.', but I'd prefer to use the value which is returned by git describe without modifications.

As @Shreeshrii already said, 59 is the number of commits since the last tagged version. The 'g' before the hash might stand for 'git'.