I’m playing with evaluation copy of BP-VA AE and I’ve found a pair of warts related to project publishing in batch mode.
First, all text strings in generated HTML files are found to be “platform encoded” (cp1251 in my case) - in spite of META directive which (fairly and hopefully) declares UTF-8 encoding.
Second, my local translations are not honored during batch mode publishing. Actually, the language file (in UserLanguage folder) is fully ignored in this mode.
Thank you for your post. I’m Lilian, a colleague of Jick.
We’ve tried to repeat your problem but not successful. Could you send me a screeshot for the encoding problem so that we can have further investigation? Thanks in advance!
Hi Lilian,
Thank you for your quick response.
I'm attaching two screenshots as well as their sources generated by BP-VA. Please note that IDE-generated screen contains translated names of model elements and subheadings according to my .lng file.
cmd_publish.zip: results from command-line publishing.
IDE_publish.zip: results via IDE.
Nad
``` <br><a class='attachment' href='/uploads/default/original/2X/6/6c285db7e7cf4a80ca02a5fdac4fb8ae5ce6de4c.zip'>IDE_publish.zip</a> <br><a class='attachment' href='/uploads/default/original/2X/6/6b23285a2dc42f591e306affc81dc9020e006bc8.zip'>cmd_publish.zip</a>
Thank you, but I’d like to confess that I’m slightly discouraged.
VP-Suite is an absolute overkill for our needs, and we were talking about BP-VA, weren’t we?
Nevertheless, I tried this VP-Suite patch:
Yes, the language translations are now honored during project
publishing in batch mode.
But no, the resulting HTML files are not correct:
they are still encoded in Windows ANSI code page in spite of the META
directive that declares UTF-8 encoding.
Thank you for trying out the fix. We thought you were installing VP Suite (VP Suite contains BP-VA).That’s why we sent you VP Suite. Sorry about this. We will send you BP-VA from now on.
Let’s go back to your question first.
Could you send us the following files for further investigation?
a screenshot of your BP-VA About dialog box
the script you used to run ProjectPublisher.bat
the ProjectPublisher.bat file
a screenshot of your ‘Regional and Language Options’ dialog box
Finally, does the problem occur ONLY IN BATCH MODE? This is very important.
Yes, the problem occurs ONLY IN BATCH MODE, as I’ve clearly stated before.
Now I’m attaching the whole bunch of data, and I hope it will help your devs. Let me explain in detail.
I created the fresh workspace in my D:\Temp directory and created there a new simple project with cyrillic name “Тест” (“Test” in English). Then I published the project into “publish_from_ide” subfolder, exited from BP-VA and published the project from command line into “publish_from_cmd” subfolder. I’ve run the ProjectPublisher.bat file manually (see publish.jpg screenshot).
Finally, I zipped all the contents of Temp folder and now I’m attaching it here.
As I’ve said earlier, the latest VP-Suite patch solved one of 2 problems. I hope these corrections will be applied to BP_VA too.
Thank you for providing the details. Would you mind trying the latest patch of BP-VA to see whether the problem still exists? The latest patch contains more fixes on generating reports and project publisher, and our engineers suspect your problem has already being fixed in the patch.
Please download our latest patch in the following link. If there is anything I can help, please do not hesitate to contact me.
The last patch has finally solved the problems.
Thank you so much!
(BTW, I’ve had one more related question, which is now obsolete:
setting aside i18n, why on earth HTML files generated in both modes were
different attribute-wise? That was an obvious sign of alert for me.)
Thank you again,
quite-impatient-for-remaining-problems-to-be-solved-ly,
They are supposed to be the same. The problem was due to the publishing process look for a class from a jar file which wasn’t present in the class path written in the script, which resulted in looking from the JRE. And regretfully, the one supplied by the JRE have problem (the encoding problem you experienced), that’s why the output generated from both modes were different. The patch fixed the problem script by adding the required jar filepath in classpath, so the output published by both modes are now the same.