QAAutomationIRCMeetings-1

This is the summarized IRC-protocol of the first Automation meetinbg on IRC found place at Monday 9th July

• • • • BEGIN LOGGING AT Mon *Jul 9 10:29:42 2007

1. Time shift of IRC meeting possible (-2 hrs for China) ? (All/HDE)

• Jul 09 11:00:59 xHDE I welcome you all to the first QA Automation IRC Meeting on this channel. Jogi has reported sick today (get well soon) and we decided not to cancel the meeting but let me take the lead instead. For those who don't know me by now my Name is Helge Delfs and I'm working as a QA Engineer in SUN QA Automation team testing Word Processor with testtool, developing QA tooling and I'm the OOo QA Automation Team Lead. So it is possible that you
• Jul 09 11:00:59 xHDE will get in contact with me in the future :-) My first target was to bring communication between SUN Automation and OOo QA Automation more together...that's why we decided to make this meeting possible.....have fun  !
• Jul 09 11:00:59 xHDE As previously announced on dev@qa.openoffice.org this meeting will be locked and posted to Ooo-Wiki afterwards:
• Jul 09 11:00:59 xHDE http://wiki.services.openoffice.org/wiki/QAAutomationIRCMeetings
• Jul 09 11:02:29 xHDE First point on agenda:
• Jul 09 11:02:29 xHDE Time shift of IRC meeting possible (-2 hrs for China) ?
• Jul 09 11:02:29 xHDE From my point of view it should be possible to attend this meeting for all people from different countries interested in QA Automation. Vito Smolej made a suggestion on mailinglist to start this meeting on 2PM GMT. Most of all from different countrys could attend. What do you think ?
• Jul 09 11:03:02 fha Ok for me.
• Jul 09 11:03:17 skotti No objections
• Jul 09 11:03:19 bosi 1 hour tooo early for me
• Jul 09 11:03:23 oliverc OK for me too
• Jul 09 11:04:09 bosi would that be 9 european time?
• Jul 09 11:11:07 xHDE So lets come back to Jogi's proposal to start at 9 a.m. MEST
• Jul 09 11:11:16 xHDE Most agreed.....
• Jul 09 11:19:22 xHDE Ok, closed. Next meeting on 09:00 a.m. MEST / 7.00 UTC Monday 16th of Juli
• Jul 09 11:19:42 sun_msc OK, agreed

2. Experiences with IRC after 1st week

• Jul 09 11:19:57 xHDE Second point: Experiences with IRC after 1st week
• Jul 09 11:19:58 xHDE From my point of view the communication within the team has been increased and is visible to all team members. Surely we need more time to accustom with it. What I miss is an extra notification if new personal messages arrive. We use X-Chat 2.4.5 on Solaris. If one has another working client as a tip feel free to post.
• Jul 09 11:20:57 shaunmcdonald xHDE: I'm using Colloquy on the Mac, which has settings for the notifications
• Jul 09 11:21:28 skotti xHDE: X-Chat can beep on new messages on SunRay. Not perfect, but good enough.
• Jul 09 11:21:53 ExLachs_home you can even filter foir buzzwords to beep only
• Jul 09 11:23:05 skotti shaunmcdonald: Sun QA is exclusively using the SunRay for their work. This somewhat limits our choice in software. But i think X-Chat is ok. We've got Gaim installed as well
• Jul 09 11:23:35 skotti If someone needs an alternative he might try it out. I did not get too happy with it.
• Jul 09 11:23:37 * shaunmcdonald has seen it in action
• Jul 09 11:24:05 * shaunmcdonald doesn't know if it sends sound
• Jul 09 11:24:20 shaunmcdonald skotti: it could be a sound problem
• Jul 09 11:24:34 xHDE Any experiences with Firefox chatzilla...should be platform independent !?
• Jul 09 11:24:39 skotti shaunmcdonald: It does (if not deliberately muted by the user)
• Jul 09 11:25:07 bosi xHDE, chatzilla: the window jumps to the front - you can't miss it
• Jul 09 11:25:21 xHDE Sound could be a problem if you have a nervous neighbour ;-)
• Jul 09 11:25:50 skotti xHDE: Certainly, you should watch bosi when my SunRay beeps. He's just soo jumpy... ;-)
• Jul 09 11:26:13 bosi skotti, bh bh ba
• Jul 09 11:26:32 shaunmcdonald What's the next item on the agenda?

3. Which TT version has to be used  ?

• Jul 09 11:27:19 xHDE Therfor I hand over to TBO to give a statement which testtool has to be used to test current builds on different platforms until further notice. Please....
• Jul 09 11:27:50 bosi usually you use the old src680m50 in any case except:
• Jul 09 11:28:52 skotti Just a remark: This is StarOffice only. For OOo please use the Testtool that comes with the application
• Jul 09 11:29:14 bosi 1. you are on solaris 10 or 11 where you have the to use a newer one, like the current src680m219 which has other problems.
• Jul 09 11:29:43 shaunmcdonald bosi: I thought that at the start of the year there was a significant change in the testtool environment, which meant that a more recent testtool was needed
• Jul 09 11:35:28 bosi but you should all keep in mind, that using the latest and greatest is not ok in software testing
• Jul 09 11:53:41 xHDE Again: Outcome of 3rd point was to use testtool m50 on all platforms except for Solaris 10 and 11. Mac uses the latest and greatest. Agreed ?
• Jul 09 12:01:05 shaunmcdonald fha: we need to specify the platforms that we know testtool doesn't work
• Jul 09 12:02:43 shaunmcdonald is there a page already to download the current testtools?
• Jul 09 12:03:46 xHDE shaunmcdonald: What about http://qa.openoffice.org/ooQAReloaded/AutomationTeamsite/ooQA-TeamAutomationBin.html
• Jul 09 12:04:29 shaunmcdonald xHDE: so we already have a page that specifys the Sun supported OSs
• Jul 09 12:04:56 xHDE shaunmcdonald: Jupp :-)
• Jul 09 12:05:43 skotti shaunmcdonald: The lists do not match, you can - without problems - use the VCL Testtool on non-supported platforms. The support-limitation is provided by system installer and glibc.
• Jul 09 12:05:52 shaunmcdonald xHDE: should we include the Sun unsupported platforms on that page?
• Jul 09 12:06:26 xHDE shaunmcdonald: If wnat to have a up-to-date list I think it makes sense
• Jul 09 12:06:48 shaunmcdonald wouldn't it be better to have the 2.0 testtool before the 1.1 testtool?
• Jul 09 12:08:00 shaunmcdonald having the most recent first would be better so that you can easier see the one which the user wants to use
• Jul 09 12:08:10 xHDE agreed
• Jul 09 12:08:26 shaunmcdonald Also the headings "VCL TestTool Binaries OOo 2.0" are very difficult to see
• Jul 09 12:09:19 skotti shaunmcdonald: Make it "2.x"
• Jul 09 12:09:32 shaunmcdonald that's another good point
• Jul 09 12:09:52 xHDE Ok, I'll fix that !
• Jul 09 12:19:44 xHDE and I will update the Testtool-Binariy pages
• Jul 09 12:19:58 xHDE dissenting votes ?
• Jul 09 12:21:18 shaunmcdonald xHDE: Mac OS X/All variations cannot use m50
• Jul 09 12:21:44 xHDE shaunmcdonald: I think this is clear now :-)
• Jul 09 12:22:13 * johanhenselmans_ has quit (Read error: 110 (Connection timed out))
• Jul 09 12:22:27 xHDE 3th poin closed
• Jul 09 12:22:30 xHDE point

4th point: Get rid-off "First doc!" window? (JSI)

• Jul 09 12:22:37 xHDE 4th point: Get rid-off "First doc!" window? (JSI)
• Jul 09 12:22:37 xHDE As JSI is sick and I can only assume what he has planned I recommend to move this point to a later meeting. Any doubts about this ? Or has someone more information on what he has in mind about this in detail ?
• Jul 09 12:24:01 shaunmcdonald the "First doc!" window is the window that is opened at the start of every test ?
• Jul 09 12:24:08 fha Yes
• Jul 09 12:25:32 shaunmcdonald it would probably be a good thing to remove it, if there is no useful purpose, as it would then give a performance boost and reduce the time it takes to do the tests
• Jul 09 12:26:43 fha Not to mention the reduced possibility of a focusing-problem...
• Jul 09 12:29:17 skotti Let's postpone the subject, i would like to hear JSI's take on the matter. He obviously wants to remove the window but i would like to get the reasons first hand. This would give me a little more time to evaluate the impact on the framework module too.
• Jul 09 12:31:23 shaunmcdonald skotti: I think postponing this item would be a good idea
• Jul 09 12:32:05 fha skotti: why not just make a sub which creates that state, before the test with that requirement starts?
• Jul 09 12:32:27 skotti fha: This is an option. Yes. Later.#
• Jul 09 12:32:49 fha skotti: would of course prolong the testing, but it could also just be something which buys you time to change your code...
• Jul 09 12:33:43 fha xHDE: shall we then postpone this item?
• Jul 09 12:39:15 shaunmcdonald fha: assume yes
• Jul 09 12:44:12 skotti We close this topic.
• Jul 09 12:44:17 skotti What's next?

5. Other

• Jul 09 12:44:56 skotti sun_msc: You had something about usage of variables like i, j, k
• Jul 09 12:45:23 shaunmcdonald skotti: the good old single letter variables :-)
• Jul 09 12:45:48 skotti shaunmcdonald: Still to be found in many places of the older testscripts.
• Jul 09 12:46:16 skotti I guess we all agree that they should be avoided whenever possible - which means: Always.
• Jul 09 12:46:33 shaunmcdonald skotti: they are still common practice even in newer books and university lectures, especially for loops
• Jul 09 12:47:26 fha skotti: For counters, I'd still like to use them sometimes.
• Jul 09 12:48:08 skotti fha: Ok, might be fine when using "small" loops. As soon as you have nested loops they become a PITA
• Jul 09 12:48:53 shaunmcdonald skotti: agreed, but it can be difficult to come up with some decent names for them
• Jul 09 12:49:02 skotti After all - what's so bad about writing iCurrentDocument instead of just "i"?
• Jul 09 12:49:18 skotti shaunmcdonald: No, that's usually fairly easy.
• Jul 09 12:49:21 fha shaunmcdonald; that's exactly what it is.
• Jul 09 12:49:53 skotti iCurrentRow, iCurrentCell, iCurrentDataSet ...
• Jul 09 12:50:29 shaunmcdonald skotti: that makes sense
• Jul 09 12:51:07 skotti You should not forget that others should have a chance to read your code as well. When using speaking names everyone can instantly see what the loop is about.
• Jul 09 12:51:10 * shaunmcdonald wonders why he just spent the last 4 years at university getting himself into debt when he learns so much more direct from the industry
• Jul 09 12:51:16 fha skotti: still, a loop inside iCurrentRow, which in that part never occurs again, could do just fine with i.
• Jul 09 12:51:43 skotti fha: But what are you doing within iCurrentRow?
• Jul 09 12:51:58 skotti iInserCurrentText?
• Jul 09 12:52:10 skotti iDeleteCurrentItem?
• Jul 09 12:52:18 * maho has quit (Read error: 104 (Connection reset by peer))
• Jul 09 12:52:38 shaunmcdonald skotti: going through each item in the row?
• Jul 09 12:52:41 * maho (n=maho@rikad85.riken.jp) has joined #qa.openoffice.org
• Jul 09 12:52:42 fha skotti: Just an example for a big loop with some smaller ones inside them.
• Jul 09 12:53:05 skotti shaunmcdonald: The first thing you can learn here is that - whether we come to a conclusion or not - everyone will just continue doing what he's used to, no matter what we agreed on ;-)
• Jul 09 12:53:31 shaunmcdonald skotti: old habits die hard ;-)
• Jul 09 12:53:48 * johanhenselmans has quit (Read error: 110 (Connection timed out))
• Jul 09 12:53:53 fha shaunmcdonald: depends...
• Jul 09 12:54:48 skotti fha: The more lopps you mix the harder it gets for somebody else to fiddle out what just this special i does differently than the i from before - or was it k?
• Jul 09 12:54:57 skotti ^lopps^loops
• Jul 09 12:55:45 fha shaunmcdonald: if taking responsibility means to change, and the change means improved code... and the improvement is beneficial for the writer... then it's motivating.
• Jul 09 12:56:00 skotti Short variables were a decent way of saving sace in the old days of 64K, nowadays our main goal should be maintainability and readability especially with the community in mind.
• Jul 09 12:57:07 fha skotti: trust me, you don't want to debug i_pEngine -test. And it contains nearly only really big loops.
• Jul 09 12:57:08 shaunmcdonald fha: anything that helps the writer is always motivating as long as it doesn't take too much longer
• Jul 09 12:57:09 skotti so people, give your votes - who's for generally using speaking variables
• Jul 09 12:57:47 shaunmcdonald skotti: what are the other options?
• Jul 09 12:57:55 shaunmcdonald in the vote?
• Jul 09 12:58:13 skotti shaunmcdonald: That would be "Do as you like", i'm afraid.
• Jul 09 12:58:15 fha skotti: I'm for.
• Jul 09 12:58:49 fha skotti: I'm for "do as you like" as well.
• Jul 09 12:58:51 bosi long speaking names, if voted here it would be mandatory for all new tests.
• Jul 09 12:59:38 shaunmcdonald speaking variables would be best practice
• Jul 09 12:59:42 fha Freedom of choice is valuable...
• Jul 09 12:59:48 skotti Ok, again: make it mandatory to use speaking variable names? Yes or no.
• Jul 09 13:00:23 oliverc yes
• Jul 09 13:00:24 fha skotti: No. Not mandatory.
• Jul 09 13:00:38 skotti yes from me
• Jul 09 13:01:24 skotti Anyone else?
• Jul 09 13:01:47 bosi yes
• Jul 09 13:01:50 shaunmcdonald how about highly recommended, with code that has one or 2 tight loops using letter variables to be allowed, however code that only has single letter variables should be not allowed
• Jul 09 13:02:09 * shaunmcdonald is just trying to be complicated :-(
• Jul 09 13:02:09 skotti shaunmcdonald: No, that's no good idea. Keep it straight.
• Jul 09 13:02:30 skotti Ok, that's it then. Using speaking variable names is now mandatory.
• Jul 09 13:02:43 skotti Other subjects to discuss?
• Jul 09 13:03:22 shaunmcdonald the date and time of the next meeting was set at the start of the meeting
• Jul 09 13:03:31 shaunmcdonald what about the log of the meeting?
• Jul 09 13:03:37 sun_msc yes for speaking variable names
• Jul 09 13:04:11 skotti Ok, Next meeting on 09:00 a.m. MEST / 7.00 UTC
• Jul 09 13:04:30 shaunmcdonald skotti: which day?
• Jul 09 13:04:39 skotti HDE will post this meeting into Wiki, exact location should be provided by him when done
• Jul 09 13:04:52 skotti shaunmcdonald: That'll be next monday.
• Jul 09 13:05:10 skotti The meeting is closed.
• Jul 09 13:17:24 skotti xHDE: I took over, the meeting was closed at 13:05
• Jul 09 13:17:38 xHDE skotti: Ok, thanks
• Jul 09 13:53:26 xHDE I'm happy to announce the first person joining the OOo Automation Team - Andre Schnabel from the german OOo community. Andre has already done some work for the localization of automated tests regarding Vietnamese language. Please welcome Andre Schnabel as a further contact person for localisation and automation of OOo.