I searched back a bit and haven't seen this asked or answered, so here goes:
I need to get the info from the .res file (the entire .res file) in ASCII format (non-binary). I'm starting on a command line utility to parse through the .res and spit out straight ASCII text but if a newer version of QAP/SilkTest does this we will invest in that rather than spending my time on it. I'm not looking for help with the coding, just news on the availability of something to do this.
We're running SilkTest 5.0 (win32). I'm not sure if Segue has improved the ResExportOnClose() stuff at all since then. Any help is appreciated, thanks!
SW Engineer / QAP Stud :-D
I personally don't have much need for the res file from Silk. In fact if there was a way for me to turn it off I would. I find these files get curupted very easilly and
cannot be read DURING a test run.
What I have done (and there are others like me) is simply log files to a shared drive. Then I support logging through a EventClass that really tells me in ASCII the important stuff.
Like did a critrical step pass or fail, how long it took etc. I even have error counts in case you detect something bad and you can log an error. This is a billion times easier than trying to reverse engineer Segue's proprietary results files.
I have a simple nameing convention that logs files by:
So that even two tests running at the same time on different machines won't collide. The choice is yours and you may not like the solution. I find thsi the only sane alternative to the bulky res file that Segue likes so very much. Besides you can read this thing with a simple editor WHILE a test is running.
Thanks Rick, I've thought about doing it that way. It wouldn't be hard to wrap the log writes so that the output is forked between the log and another .txt file. I'm just trying to use what's there already and I guess I should give up on that :-D. The QAP stuff that we are running in my department was all written by me so it won't be hard to go back and add more things in. Thanks for the suggestion/reinforcement that the Segue-way isn't the best.
I have never been able to figure out the (compressed?) binary format of the ".res" files.
The easiest way to convert these to ascii is to select "Export Results". This is the same in SilkTest 5.0 as in 5.03.1.
I suspect that you want to automate this process. If so, check out the ResExport() and ResExportOnClose() functions that you can easily add to your test.
I posted this (after being interupted) and I see rick_weth's post.
I agree with him to a large extent, and I have used my own results output as well.
Segue's results viewer has some nice features that I like (like the difference viewer). But do hate the impenetrable format.
[This message has been edited by RobKapteyn (edited 11-21-2000).]
I almost have the .res format figgered ... it's 'impenetrable' to me because the header contains variable lengths of info, with no reasonable explanation in sight. Some of the fields are fairly obvious; it contains a TOC of sorts to the offset of the beginning of each set of results. If you do an 'export', look at the #fields line to see what is in each header space.
Yes, though, you hit the real problem on the head, and that is that it needs to be automated. I'm going with the 'rewrite the log process' approach. Everything that we need to know for reporting is accessible via the "Script Information Functions" provided with Silk. At least this way we have total control over the output, and can skip the intermediate step and put it straight into a TET compliant format (e.g. 31|0|0| "Testcase name" | platform, etc.).