Chris
--
C J Craig
Ch...@skipton.demon.co.uk
Iyonix ARM XScale computer Risc OS 5.14
Why? Do think think such comments help?
In the real world, bug reports need to be detailed with exact messages.
Moreover, just because it's been reported doesn't mean anyone has
the motivation, time or skill to fix it.
> C J Craig wrote:
>> Firefox works well on my Iyonix. More often than not when I load
>> Firefox. No Sprite appears on the icon bar. Indeed after asking if I
>> want a new session it reports the Sprite is missing.
<snip>
> In the real world, bug reports need to be detailed with exact messages.
> Moreover, just because it's been reported doesn't mean anyone has
> the motivation, time or skill to fix it.
The problem appears to be caused when Firefox hasn't been seen by the
filer before being run. This happened on my system when I select
Firefox from the Tasks menu of !DialUp and !NetFetch.
The solution is a simple matter of adding the following line after the
lines that set Firefox$Dir and Firefox$Path:-
Iconsprites <Firefox$Dir>.!Sprites
Shouldn't take more than a couple of seconds. ;-)
How are you getting on with certificates, by the way? It's the only
real problem I have with using Firefox.
--
Graham Thurlwell
Jades' First Encounters Site
http://www.jades.org/ffe.htm
Yes, perhaps. But something no one bothered to tell me about until now.
> How are you getting on with certificates, by the way? It's the only
> real problem I have with using Firefox.
Probably be done when the 3 conditions I named above are met. It's
certainly not the only issue, and no one came forward on the
most recent request for someone to organize bug reports and features,
and there's definitely no one obviously interested in development.
Anyway, certificates mean shared libraries, which probably means
RiscPkg deployment, which is going to cause far more drama than
I care to deal with right now.
> Graham Thurlwell wrote:
>> On the 24 Nov 2009, Peter Naulls <pe...@chocky.org> wrote:
>>
>>
>>> In the real world, bug reports need to be detailed with exact messages.
>>> Moreover, just because it's been reported doesn't mean anyone has
>>> the motivation, time or skill to fix it.
>>
>> The problem appears to be caused when Firefox hasn't been seen by the
>> filer before being run. This happened on my system when I select
>> Firefox from the Tasks menu of !DialUp and !NetFetch.
>>
>> The solution is a simple matter of adding the following line after the
>> lines that set Firefox$Dir and Firefox$Path:-
>>
>> Iconsprites <Firefox$Dir>.!Sprites
> Yes, perhaps. But something no one bothered to tell me about until now.
Pardon, Sir, error! I did post on just this subject, and gave the
solution, several years ago. I can't find it in my archives (I think I
was using Pluto then, and the files aren't easily accessible), but I
vividly remember being severely ticked-off for mentioning the subject.
Since then, whenever the subject has arisen, I've posted replies
off-list.
[snip]
With best wishes,
Peter.
--
Peter, \ / \ Prestbury, Cheltenham, Glos. GL52
Anne \ / __ __ \ England.
and / / \ | | |\ | / _ \ http://pnyoung.orpheusweb.co.uk
family / \__/ \_/ | \| \__/ \______________ pny...@ormail.co.uk.
Let's tone down the drama, yes?
I did post on just this subject, and gave the
> solution, several years ago. I can't find it in my archives (I think I
> was using Pluto then, and the files aren't easily accessible), but I
> vividly remember being severely ticked-off for mentioning the subject.
If you say so. But this returns to exactly the point which you've
nicely snipped, which is about needing to collect proper bug reports,
and ensure they're dealt with. Otherwise, they go on forever, and
are never addressed. Certainly I cannot remember 100% of the issues.
> Since then, whenever the subject has arisen, I've posted replies
> off-list.
Fantastic. So very RISC OS, one off hacky solutions, and doing stuff
without bothering to collaborate - and I'm again forced to return
to PV's nonsense about doing his own (wrong) Firefox instructions
without bothering to consult me. *This* is one of the things
that makes RISC OS development so very tedious. And please, no
remonstrations about not being a programmer, etc, etc.
When you say that, do you mean:
a) you posted about it on the csa* newsgroups and mentioned the
workaround
or
b) you raised a bug report with Peter and provided the necessary patch
file to implement the workaround
Because Peter's complaint is that while plenty of people are willing
to do (a), no one ever seems to bother with even the first part of
(b), *which would be much more useful to him or anyone else working
on the port*.
Lesson 1: if no one files bug reports, no one supplies patches.
Lesson 2: if no one provides patches, bugs don't get fixed.
Chris.
> Dr Peter Young <pny...@ormail.co.uk> wrote:
>>On 25 Nov 2009 Peter Naulls <pe...@chocky.org> wrote:
>>> Graham Thurlwell wrote:
> [snip]
>>>> The solution is a simple matter of adding the following line after the
>>>> lines that set Firefox$Dir and Firefox$Path:-
>>>>
>>>> Iconsprites <Firefox$Dir>.!Sprites
>>
>>> Yes, perhaps. But something no one bothered to tell me about until now.
>>
>>Pardon, Sir, error! I did post on just this subject, and gave the
>>solution, several years ago. I can't find it in my archives (I think I
> When you say that, do you mean:
> a) you posted about it on the csa* newsgroups and mentioned the
> workaround
> or
> b) you raised a bug report with Peter and provided the necessary patch
> file to implement the workaround
The former, for reasons given below, which seemed to me to be
persuasive.
> Because Peter's complaint is that while plenty of people are willing
> to do (a), no one ever seems to bother with even the first part of
> (b), *which would be much more useful to him or anyone else working
> on the port*.
(b) seems at first sight to be sensible. However, in the Firefox port
Help file (yes, I do RTFM) I find this:
[quote]
Reporting bugs
--------------
I would much prefer you *didn't* report mis-features, things you'd
like to see, and any odd behaviour. This might seem unhelpful, but
chances are, I already know all about it, and extra emails will not
help.
I would like to hear from you only if you have a crash that is
repeatable. Please send me !Firefox.stderr and precisely what sequence
of actions is required to repeat it.
[unquote]
The absence of an icon-bar icon is at worst a minor inconvenience to
those who expect things to work in an expected way. In no way is it
the sort of crash referred to in the above quotation, so I thought I
was obeying Mr Naulls's wishes in not reporting this problem and
solution to him. After the telling-off after my first posting to a
newsgroup about this, it seemed that the best way of helping
*individuals* without arousing ire would be to post off-group.
So, please. do we report all bugs, as now seems to be the required
thing, or do we heed the instructions in the help file?
Having said that, I have no wish to drag this out, so I will now crawl
back under my stone and try to shut up.
> Firefox works well on my Iyonix. More often than not when I load
> Firefox. No Sprite appears on the icon bar. Indeed after asking if I
> want a new session it reports the Sprite is missing. Is there a simple
> answer to this? I hardly dare ask in case I invoke the wrath of P.N
> Chris
I have to admit, the threaded version doesn't work at all on my
machine. Completely locks up after look at one or two pages.
I haven't reported it to Peter, because apart from the fact it
crashes, I can tell him nothing more as the error file is always
blank, after rebooting.
I had hoped to see more people saying it did/didn't work, however it
seems I'm in a minority with the problems.
regards
--
_________________________________________
| RISC OS British Technology
| mailto:jo...@jpeachey.co.uk
_______________________| http://www.jpeachey.co.uk
> In message <b70970bf...@skipton.demon.co.uk>
> C J Craig <Ch...@skipton.demon.co.uk> wrote:
>> Firefox works well on my Iyonix. More often than not when I load
>> Firefox. No Sprite appears on the icon bar. Indeed after asking if I
>> want a new session it reports the Sprite is missing. Is there a simple
>> answer to this? I hardly dare ask in case I invoke the wrath of P.N
>> Chris
> I have to admit, the threaded version doesn't work at all on my
> machine. Completely locks up after look at one or two pages.
> I haven't reported it to Peter, because apart from the fact it
> crashes, I can tell him nothing more as the error file is always
> blank, after rebooting.
> I had hoped to see more people saying it did/didn't work, however it
> seems I'm in a minority with the problems.
> regards
If you have read the instructions & updated !Tinct etc contact Peter
Which wasn't the last word on the matter by a long shot, as I think
you know.
> After the telling-off after my first posting to a
> newsgroup about this, it seemed that the best way of helping
> *individuals* without arousing ire would be to post off-group.
>
> So, please. do we report all bugs, as now seems to be the required
> thing, or do we heed the instructions in the help file?
>
> Having said that, I have no wish to drag this out, so I will now crawl
> back under my stone and try to shut up.
Yes, well. The problem is that this passive-aggressive behaviour is
to no one's advantage. Hiding problems and postulating short-term
fixes only compounds existing problems, and is a good way to ensure
they might never be fixed properly. Unfortunately, the RISC OS way
is instead to dwell on past misdeeds (imagined or otherwise) and point
fingers, rather than trying to take ownership of problems, whether you
created them or not - i.e, showing a degree of maturity.
This isn't about the minor bug of a missing icon, but about reporting
problems in general. Ideally, such a problem would be reported in
a bug database, so that (a) a developer can know about it (b) a
developer can remember to fix it (c) its closure can be effectively
reported (d) can be used as part of general development goalposts
to indicate project completion. I appreciate that many users probably
don't know how to write a good bug report, but I do know that *most* can
be trained to do so with not much effort.
No, I don't want an apology, but it might be nice to see someone step
up with some effort once in a while.
>>
>> This isn't about the minor bug of a missing icon, but about reporting
>> problems in general. Ideally, such a problem would be reported in
>> a bug database, so that (a) a developer can know about it (b) a
>> developer can remember to fix it (c) its closure can be effectively
>> reported (d) can be used as part of general development goalposts
>> to indicate project completion. I appreciate that many users probably
>> don't know how to write a good bug report, but I do know that *most* can
>> be trained to do so with not much effort.
>> No, I don't want an apology, but it might be nice to see someone step
>> up with some effort once in a while.
I am sorry.
What a brilliant idea! How do us punters do that?
My version of Firefox will not print.
I used to have the same trouble learning French.
I remember asking my teacher what the French for "Table" was.
He always said "Well, Christopher what do you think it is?"
I replied you are my teacher, please tell me what it is?
I never did find out!
I don't think he could see the problem from my perspective.
Thanks in anticipation
> In message <347d8ac2...@skipton.demon.co.uk>
> C J Craig <Ch...@skipton.demon.co.uk> wrote:
> [snip]
>
> >>
> >> This isn't about the minor bug of a missing icon, but about reporting
> >> problems in general. Ideally, such a problem would be reported in
> >> a bug database, so that (a) a developer can know about it (b) a
> >> developer can remember to fix it (c) its closure can be effectively
> >> reported (d) can be used as part of general development goalposts
> >> to indicate project completion. I appreciate that many users probably
> >> don't know how to write a good bug report, but I do know that *most* can
> >> be trained to do so with not much effort.
>
> >> No, I don't want an apology, but it might be nice to see someone step
> >> up with some effort once in a while.
>
> I am sorry.
> What a brilliant idea! How do us punters do that?
> My version of Firefox will not print.
>
> I used to have the same trouble learning French.
> I remember asking my teacher what the French for "Table" was.
> He always said "Well, Christopher what do you think it is?"
> I replied you are my teacher, please tell me what it is?
Here's a suggestion. You know the English for table. Why not try to guess at
the French equivalent? You know, take a wild stab in the dark? If you truly
don't have enough initiative to even try a guess, (which is what the teacher
was trying to teach you, initiative), then you're beyond help, frankly.
I've spent a good few years working in IT support, and the users who refuse
to try anything unless they're having their hands held are by far the
biggest time sinks. I won't help that sort of person unless I'm being paid
to.
I have HAD technical conversations like this:
<talking a user through doing something>
...
Them "Now what do I do?"
Me: "What does it say on the screen?"
Them: "It says, 'Click OK to continue.'"
Me: "So ...?"
Them: "Shall I click OK?"
Me: "Do you want to continue?"
Them: "I don't know."
...
They're soul-destroying. They make you wonder how the user concerned even
manages to continue breathing without someone to remind them. If you truly
are one of these people, for god's sake let us know and I for one will
killfile you ASAP. The alternative is likely to involve a light spot of axe
murder.
> I never did find out!
Why the f* not?! Why should /anyone/ help you if you can't even solve a
problem this simple without assistance? Certainly I won't be bothering
again. If discovering the French for table is beyond you, why should anyone
trust you to read readme files? Bother with bug reports? Follow installation
instructions? I mean, how did you ever manage to download Firefox in the
first place if no-one ever told you to? Or did someone else do it for you?
> I don't think he could see the problem from my perspective.
Can you see what /he/ was trying to teach you? After my hint about about
having a guess?
And people wonder why IT techs are such heavy drinkers...
--
Simon Smith
When emailing me, please use my preferred email address, which is on my web
site at http://www.simon-smith.org
> In message <4bee8bc2...@skipton.demon.co.uk>
> C J Craig <Ch...@skipton.demon.co.uk> wrote:
>> In message <347d8ac2...@skipton.demon.co.uk>
>> C J Craig <Ch...@skipton.demon.co.uk> wrote:
>> [snip]
>>
>>>>
>>>> This isn't about the minor bug of a missing icon, but about reporting
[snip]
>>
>> I am sorry.
>> What a brilliant idea! How do us punters do that?
>> My version of Firefox will not print.
>>
>> I used to have the same trouble learning French.
>> I remember asking my teacher what the French for "Table" was.
>> He always said "Well, Christopher what do you think it is?"
>> I replied you are my teacher, please tell me what it is?
[snip]
Actually I did quite well in my O level French. I just used Emglish
words!
But why waste so much bandwidth getting in a Rant. I learn nothing
from that. We could all learn so much.
You could just say:
"Please send me !Firefox.stderr and precisely what sequence of actions
is required to repeat it."
Pity it does not quite say were to send it!
At least I managed to get your attention.
> > I used to have the same trouble learning French. I remember asking my
> > teacher what the French for "Table" was. He always said "Well,
> > Christopher what do you think it is?" I replied you are my teacher,
> > please tell me what it is?
>
> Here's a suggestion. You know the English for table. Why not try to guess
> at the French equivalent? You know, take a wild stab in the dark?
Hmm.
Is it, perhaps, "hohehoheho" ?
[...]
> I've spent a good few years working in IT support, and the users who
> refuse to try anything unless they're having their hands held are by far
> the biggest time sinks. I won't help that sort of person unless I'm being
> paid to.
Comparing IT support with teaching another language isn't comparing apples
with pears, it's comparing apples with digital watches. Yes, it was Chris
who actually brought the comparison in by likening Peter's comments with
those of his French teacher, but there's no need to buy into and further the
comparison when it just doesn't work.
[...]
--
Vince M Hudd - Soft Rock Software
http://www.softrock.co.uk
http://misc.vinceh.com
> Actually I did quite well in my O level French. I just used Emglish
> words!
>
[snip irrelevant nonsense]
> But why waste so much bandwidth getting in a Rant. I learn nothing
> from that. We could all learn so much.
We could, but it's not clear to me that people actually want to
learn things. Or "learn" things that might run counter to what
they already believe.
>
> You could just say:
>
> "Please send me !Firefox.stderr and precisely what sequence of actions
> is required to repeat it."
>
> Pity it does not quite say were to send it!
>
> At least I managed to get your attention.
Simon wouldn't say that, nor is getting his attention necessarily a
good thing, since he is not a RISC OS Firefox port developer.
Anyway, I'm not at all interested in playing these silly games with
comparisons.
As to the stderr question - that's only useful for certain types of
crashes - even then, it has to be accompanied by useful information
in terms of repeatability, etc - stuff I implied earlier. The
extrapolation here is that I then go away and do lots of work and
magically fix stuff with minimal effort on your part - and rather
contrary to the aim of learning anything at all.
Anyway, if someone is serious about properly reporting bugs (and
features) then I will certainly work with them do to so. Not just
on Firefox, but on all riscos.info ports. If nothing else, the time
saving to everyone will be significant.
> C J Craig wrote:
>> In message <ed088fc25...@zen.co.uk>
>> Simon Smith <simon_sm...@zen.co.uk> wrote:
> We could, but it's not clear to me that people actually want to
> learn things. Or "learn" things that might run counter to what
> they already believe.
>>
> Anyway, if someone is serious about properly reporting bugs (and
> features) then I will certainly work with them do to so. Not just
> on Firefox, but on all riscos.info ports. If nothing else, the time
> saving to everyone will be significant.
Yes, I am happy to have a go at that. Unfortunately the Helpfile in
!Firefox does not make it clear. Where to report bugs? Do you mean as
below?
http://www.riscos.info/index.php/GCCSDK
Can you attatch files to these post?
Many thanks for your speedy reply. I try to see thing from your shoes!
> > Anyway, if someone is serious about properly reporting bugs (and
> > features) then I will certainly work with them do to so. Not just on
> > Firefox, but on all riscos.info ports. If nothing else, the time
> > saving to everyone will be significant.
>
> Yes, I am happy to have a go at that.
Just to add another perspective to this, as it was a subject that cropped up
in conversation this evening:
If someone sends me a concise and comprehensive bug report that gives me
enough detail to reproduce the problem immediately, then I can often fix the
bug there and then. Time taken -- say -- 1 hour.
If they don't, then I have to spend time trying to guess what they might
have done, and then compose an email back guiding them towards giving me the
information that I need to be able to reproduce the issue myself. If that
takes a couple of tos and fros by email, then it may well eat up several
evenings of my time. Once the bug is finally reproduced, I still have to
spend that hour (or whatever) fixing it.
That isn't to say that I (or other developers, I assume) don't want your bug
reports. What *is* being said is that by your spending some time putting
together a useful bug report, you can save *us* time. And when RISC OS
development is increasingly a hobby, and fighting for time against --
frankly -- more interesting pastimes, helping the developers spend less time
chasing bug reports directly benefits *you* by allowing more development to
actually happen.
I've not seen anyone post this link to the thread yet, so for completeness,
here it is:
http://www.chiark.greenend.org.uk/~sgtatham/bugs.html
It's a comprehensive guide to how to report bugs, along with (hopefully) an
illustration of why it matters to get it right. It's worth a read before
you next start to write an email to a developer.
--
Steve Fryatt - Leeds, England
> I've not seen anyone post this link to the thread yet, so
> for completeness, here it is:
> http://www.chiark.greenend.org.uk/~sgtatham/bugs.html
> It's a comprehensive guide to how to report bugs, along
> with (hopefully) an illustration of why it matters to get
> it right. It's worth a read before you next start to
> write an email to a developer.
That seems, on a first read, to be a very useful article.
The one thing that it seemed to me to leave out, was a clear
list of things that arguably should always be included;
something along the lines of:
Make / model of machine
OS version
Amount of Ram
Any additional pieces of non-standard hardware
Names of utilities or modules loaded as standard
Any other things?
Would that make sense?
--
Russell
http://www.russell-hafter-holidays.co.uk
Russell Hafter Holidays E-mail to enquiries at our domain
Need a hotel? <http://www.hrs.com/?client=en__blue&customerId=416873103>
Is that backed by stats?
I thought the beer, coffee and biscuits[1] just came along with the other
slightly aberrant behaviour.
At the risk of promoting an extended argument, would it be a good idea for
one of the software bods to put together a loose pro-forma for bug reports
so that ordinary users can give the level of detail required.
Only they know how much detail that is, and Steve Fryatt has pointed out
how wasteful additional drill down questioning can be in terms of time, but
equally too much (trivial) detail may slow things down too.
--
Steve Pampling
> > And people wonder why IT techs are such heavy drinkers...
> Is that backed by stats?
No, it's backed by anything with Vodka in it.
[...]
> At the risk of promoting an extended argument, would it be a good idea for
> one of the software bods to put together a loose pro-forma for bug reports
> so that ordinary users can give the level of detail required. Only they
> know how much detail that is, and Steve Fryatt has pointed out how
> wasteful additional drill down questioning can be in terms of time, but
> equally too much (trivial) detail may slow things down too.
That's actually a very good idea, apart from one very fatal flaw. In all
likelihood, while some people will make use of it, most will just ignore it
and continue to submit bug reports that are somewhat lacking in clarity and
useful information. This may be because they'll just forget where it is, but
in some cases it will be because they know better and their bug reports are
a model of clarity and they can't understand why other people - those they
are submitting the reports to in particular - can't see that.
Believe it or not, I hadn't forgotten. Here it is:
> http://www.riscos.info/index.php?title=Special:AWCforum&action=st/id12
> 6/Bug_reporting_now_open_on_riscos.info
Many thanks, I will now try & find some bugs!
Happy New Year.