Comment #2 on issue 880 by [email protected]: Possibility to generate individual log files for each combined suite
http://code.google.com/p/robotframework/issues/detail?id=880
The main issue I have with any re-generated log file is that to date the rebot process is not intelligent about embedded HTML references. All our in-house Robot Framework test engines have a feature where they automatically capture a screenshot of the SUT and embed this file as an HTML reference in the produced log.html file. Such references look like this: <tr><td colspan="3"><a href="FAIL_Screen_1_280~01.png"><img src="FAIL_Screen_1_280~01.png" width="320px"></a></td></tr> The captured FAIL_Screen_1_280~01.png file exists in the same folder as the original log.html and output.xml files.
If this issue's proposed solution is implemented, instead of re-using existing log.html files as I proposed in Issue 876, then rebot MUST be made smart enough to recognize these "in the same folder" HTML references and correct them if it creates a new log.html file in a different folder. For example, using the sample folder structure that I attached to Issue 876, if rebot was to produce say a log1.html file at the top BAT folder level, and if this log file contained an HTML reference to a captured screen shot like the one shown above that originally came from BAT/Devices/Dev_NCAS/log.html, then rebot should correct the HTML reference like this: <tr><td colspan="3"><a href="BAT/Devices/Dev_NCAS/FAIL_Screen_1_280~01.png"><img src="BAT/Devices/Dev_NCAS/FAIL_Screen_1_280~01.png" width="320px"></a></td></tr>
Without this href=... and img src=... correction the rebot log files are not very useful to my team.
