![]() |
Hello,
When I modify the file myfunction.sci in my librairie "mylib", the update works as expected when I restard Scilab with the following lines in my .scilab file genlib(SCIHOME+"/mylib/macros") load(SCIHOME+"/mylib/macros/lib") add_help_chapter("Librairie perso",SCIHOME+"mylib/jar") I would like to do the update without Scilab exit and restart, so I do --> del_help_chapter("Librairie perso") --> clear mylib myfunction --> genlib(SCIHOME+"/mylib/macros") --> help_from_sci(SCIHOME+"/mylib/macros/myfunction.sci",SCIHOME+"/mylib/help/fr_FR"); --> xmltojar(SCIHOME"/mylib/help/fr_FR","Librairie perso","fr_FR"); --> load(SCIHOME+"/mylib/macros/lib") --> add_help_chapter("Librairie perso",SCIHOME+"mylib/jar") Excepted genlib and xmltojar, not all the code lines work all the time (some time yes and some time no). I understand nothing because I not able to reproduce "deterministically" the success, or not! So I think my code is wrong. What is the good code to update library and help after changing a file? -- Jean-Yves Baudais _______________________________________________ users mailing list [hidden email] http://lists.scilab.org/mailman/listinfo/users |
![]() |
Hello Jean-Yves, please let me forward
Samuel Gougeon's answer to your question. Best Regards, Philipp Hello Jean-Yves,
Do you get any error message that tells that this line is KO
while by default other ones are OK?
On Windows, displaying the consolebox("on") when compiling some
documentation is the best and only way i know to see where are the
problems and fix them in source files. Without this, we blindly
work. Best regards Samuel Gougeon PS: Thanks to Philipp for forwarding this answer without breaking the thread. ---------------------------------- Am Mo., 18. Jan. 2021 um 14:55 Uhr schrieb Jean-Yves Baudais <[hidden email]>: Hello, _______________________________________________ users mailing list [hidden email] http://lists.scilab.org/mailman/listinfo/users |
![]() |
Hello,
Thanks for the message. ----- Original Message ----- > Do you get any error message that tells that this line is KO while by default > other ones are OK? > Please note that the compilation of the documentation can (quite often) yield > some errors that are [ http://bugzilla.scilab.org/7255 | NOT displayed in the > Scilab console ] (but, on Windows, in the so-called consolebox()). > After such "silent" errors, some involved files [ > http://bugzilla.scilab.org/8430 | can > stay locked ] (or simply not updated), and next instructions might look as > failing. I saw that some messages are written on XTERM and not on Scilab window. Well, I did new test with a simple function in home+"/macros" function out=testadd(a,b) // To test process for lib // // Syntax // out=testlib(a,b) // Parameters // a: scalar // b: scalar // Description // Add to scalar // Author // Me // Bibliography // Nothing out=a+b; endfunction and I generate the lib "testlib". Contrary to what I wrote, that works fine. I did mistake when I used clear. It must be --> clear testlib testadd and not --> clear testlib, testadd (testadd should also be cleared if it has been called, because load("macros/lib") is not enough to replace the function already loaded in Scilab memory, I guess) So, there is no problem with the updated library though the load is fisrtly done in .scilab. I don't need to restart Scilab to use the updated library "testlib". The problem still remains for the help: --> help_from_sci("macros","help/gl_GL") help_from_sci: Reading from directory macros help_from_sci: Processing of file: macros/testadd to help/gl_GL/testadd help_from_sci: processed 1 files. ans = "" Of course I created the help/gl_GL directory before, otherwise there is an error, and testadd.xml is created (gl is for globish, english is out of my reach) --> xmltojar("help/gl_GL","My test lib","gl_GL") jar/scilab_gl_GL_help.jar is udpated, and there is also a message in the linux XTERM with the number of generated files. --> add_help_chapter("My test lib","jar") T --> help The chapter "My test lib" is in the help window with the function testadd. --> del_help_chapter("My test lib") --> help The chapter is not in the help, as expected. --> add_help_chapter("My test lib","jar") T I change the head comment and rebuilt the help: --> help_from_sci("macros","help/gl_GL") a warning mentions that the file exists, but I choose "create anyway" --> xmltojar("help/gl_GL","My test lib","gl_GL") the files are updated --> help Could not load file: /home/jbaudais/jar/scilab_gl_GL_help.jar. Please check its contents, must be a Java Help file. Error message: Could not parse Got an IOException (JAR entry scilab_gl_GL_help/jhelpset.hs not found in /home/jbaudais/jar/scilab_gl_GL_help.jar) Parsing failed for null However, I can open the file "jhelpset.hs", so the message is wrong, the entry "scilab_gl_GL_help/jhelpset.hs" exists in the file "/home/jbaudais/jar/scilab_gl_GL_help.jar", on the filesystem of course, but Scilab use its memory (stack, heap?) and not the filesystem. Maybe the "scilab_gl_GL_help.jar" is corrupted by xmltojar on the Scilab memory. What do you think? I try --> del_help_chapter("My test lib") before modification, and --> add_help_chapter("My test lib","jar") but that change nothing. I must restart scilab and --> add_help_chapter("My test lib","jar") --> help it works fine. So, I have no solution to rebuilt the help without Scilab restart, from scratch! I think the corresponding Scilab memory is not correctly updated or is corrupted somewhere (xmltojar, add_help_chapter?), so we must restart Scilab. Do you have the same behaviour? -- Jean-Yves PS : I read the discussion about profile and did test. Is there similar tool to check memory within Scilab, or do we need to use external tools as valgrind? _______________________________________________ users mailing list [hidden email] http://lists.scilab.org/mailman/listinfo/users |
![]() |
Thanks Philipp,
Now i can answer inside the recovered thread: Jean-Yves, i remember -- from another
thread -- that you are running all this with Scilab 5.5.2. Aren't
you?
Le 29/01/2021 à 18:47, Jean-Yves Baudais a écrit : .../... Well, for more than 6 years that 5.5.2 was released, a lot of
things were done, noticeably about fixing some help_from_sci bugs. _______________________________________________ users mailing list [hidden email] http://lists.scilab.org/mailman/listinfo/users |
![]() |
Le 30/01/2021 à 15:42, Samuel Gougeon a écrit :
> Thanks Philipp, > Now i can answer inside the recovered thread: > > Jean-Yves, i remember -- from another thread -- that you are running all > this with Scilab 5.5.2. Aren't you? Yes I am. I also did tests on Scilab 6.1.0, but not on my computer. > Well, for more than 6 years that 5.5.2 was released, a lot of things > were done, noticeably about fixing some help_from_sci bugs. > Have you checked fixed bugs? If you installed uman, > --> uman help_from_sci b > <http://bugzilla.scilab.org/buglist.cgi?bug_status=NEW&bug_status=ASSIGNED&bug_status=REOPENED&bug_status=RESOLVED&list_id=59325&order=resolution%2Cbug_id%20DESC&product=Scilab%20software&query_format=advanced&short_desc=%28%5E%7C%5B%5Ea-zA-Z_%5D%29%28help_from_sci%29%28%5B%5E0-9a-zA-Z_%3B%5D%7C%24%29&short_desc_type=regexp> I didn't see any corresponding bug report. Anyway, I will first update my numerical environment (OS, tools, own libraries...) as soon as possible and comeback if needed. Thanks, --Jean-Yves _______________________________________________ users mailing list [hidden email] http://lists.scilab.org/mailman/listinfo/users |
Free forum by Nabble | Edit this page |