Invalid date format(bad integer)

[cc_rider]

Whenever cumulus starts it complains via the blinking 'error' indicator that there is an 'invalid integer value'. Upon closer investigation of the dayfile.txt and mmmyyLOG.txt files, I notice that the date field is incorrect. Cumulus is writing these files as:

dayfile.txt:
23-10-13,0.0,0,00:00,6.1,00:00,6.1,00:00,1004.1,00:00,1004.1,00:00,0.0,00:00,0.3,6.1,18.3,0.8,00:00,99,00:00,99,00:00,0.00,0.0,6.1,00:00,5.0,00:00,5.0,00:00,0.0,00:00,6.1,00:00,6.0,00:00,6.0,00:00,296,12.2,0.0,,,,
24 10 13,25.6,225,14:27,3.0,23:10,8.9,15:44,999.5,14:17,1006.1,23:58,0.0,00:00,9.6,6.0,96.5,19.4,14:12,59,15:50,99,00:00,0.00,0.0,8.9,15:44,5.1,14:43,0.6,23:16,0.0,00:00,2.6,21:17,6.0,00:00,-0.3,18:49,281,12.3,0.0,,,,

oct13log.txt:
25-10-13,15:17,1.8,90,0.3,2.3,3.6,231,0.0,0.0,1007.7,10070.0,22.0,40,2.5,1.8,1.8,0.0,0,0.00,0.00,-0.6,345,0.0,270
25 10 13,15:21,6.5,91,5.1,3.2,7.2,230,0.0,0.0,999.9,10070.0,21.3,49,7.2,6.5,6.5,0.0,0,0.00,0.00,4.8,335,0.0,225

Note the date in the first entry is hyphenated: 25-10-13. Whereas the date in the second entry not. I must, at random or regular intervals, manually edit the files to change the blanks to hyphens. How does cumulus get the date from windows(7 or 8) and why is it being so consistently incorrect? Is this a bug? This symptom began about 6 months ago, and I cannot remember what occurred around that time to instigate such a behavioral change in cumulus.

Any help appreciated.

Chris
aka cc_rider

[steve]

See my reply to your previous post (although it might be better to continue the discussion in this separate thread).

[cc_rider]

See my reply to your previous post (although it might be better to continue the discussion in this separate thread).

.
Hi Steve,
Looks like it's been a long time since I've been here,.,.

I looked at the debug files.. 001, 002 etc through to 009. They are mere text files. Interestingly I found the following timestamp:
Oct 26 16:50:44.569 : Error plotting charts at line 92 of dayfile.txt: '''' is not a valid integer value

So I opened dayfile.txt at line 92:
25 05 13,27.0,270,08:16,5.0,05:16,17.8,15:19,1011.2,15:13,1013.9,07:50,3.6,14:12,0.3,12.6,91.3,13.3,08:16,32,14:40,78,02:59,0.00,0.0,17.8,15:19,15.7,15:18,1.7,05:56,0.3,14:11,3.2,05:56,3.9,12:10,-0.2,14:54,327,5.6,0.0

However, when I look in the .001 file(prior to errors occurring), it shows the following timestamp:
2013-02-14 22:26:36.783 : System Uptime: 1 day 7 hours

Notice the difference in format. Should this be considered a bug in cumulus? It is, imho, supposed to be a standard format no matter what regional settings are, yes?

If cumulus is looking at 'short date', mine is set to 'mmm dd'. I do this because on my pc's I put the task bar up on the right side of the monitor. Using 'mmm dd yy' is too long to accommodate the width. I know which year we have, so I only want 'Oct 26' to show. This may be where cumulus is choking..? Does this help?


Regards,
Chris

[steve]

However, when I look in the .001 file(prior to errors occurring), it shows the following timestamp:
2013-02-14 22:26:36.783 : System Uptime: 1 day 7 hours

Notice the difference in format. Should this be considered a bug in cumulus? It is, imho, supposed to be a standard format no matter what regional settings are, yes?

The timestamps of the entries in the diags files are written by the logging software and aren't relevant to Cumulus itself. I don't know where the logging software gets its formats from, but the fact that your files have two different formats suggests that your Windows settings have changed - and perhaps keep changing. There is a bug in Windows 7 where the formats can change between the default US settings and the user-selected settings. There is a work around to fix this described in this post: https://cumulus.hosiene.co.uk/viewtopic.php?f=2&t=2837. I also added an alternative work around to Cumulus, but that seems not to be working for you, for some reason.

If cumulus is looking at 'short date', mine is set to 'mmm dd'. I do this because on my pc's I put the task bar up on the right side of the monitor. Using 'mmm dd yy' is too long to accommodate the width. I know which year we have, so I only want 'Oct 26' to show. This may be where cumulus is choking..? Does this help?

The date format in the log files is day<separator>month<separator>year, where <separator> is taken from Windows. The code in Cumulus actually specifies "/". I didn't know until other people started using Cumulus that specifying "/" actually means "whatever the Windows date separator is set to". I believe the separator does come from the short date format, and a space causes a problem in the system routines which Cumulus uses which parse the lines in the logs, as they treat a space character as a field separator. My understanding is that you are expected to use one of the standard separators - "/", "-" or ".". But whether it's valid or not to use a space, the system routines used by Cumulus can't cope with it.

[cc_rider]

However, when I look in the .001 file(prior to errors occurring), it shows the following timestamp:
2013-02-14 22:26:36.783 : System Uptime: 1 day 7 hours

Notice the difference in format. Should this be considered a bug in cumulus? It is, imho, supposed to be a standard format no matter what regional settings are, yes?

The timestamps of the entries in the diags files are written by the logging software and aren't relevant to Cumulus itself. I don't know where the logging software gets its formats from, but the fact that your files have two different formats suggests that your Windows settings have changed - and perhaps keep changing. There is a bug in Windows 7 where the formats can change between the default US settings and the user-selected settings. There is a work around to fix this described in this post: https://cumulus.hosiene.co.uk/viewtopic.php?f=2&t=2837. I also added an alternative work around to Cumulus, but that seems not to be working for you, for some reason.

If cumulus is looking at 'short date', mine is set to 'mmm dd'. I do this because on my pc's I put the task bar up on the right side of the monitor. Using 'mmm dd yy' is too long to accommodate the width. I know which year we have, so I only want 'Oct 26' to show. This may be where cumulus is choking..? Does this help?

The date format in the log files is day<separator>month<separator>year, where <separator> is taken from Windows. The code in Cumulus actually specifies "/". I didn't know until other people started using Cumulus that specifying "/" actually means "whatever the Windows date separator is set to". I believe the separator does come from the short date format, and a space causes a problem in the system routines which Cumulus uses which parse the lines in the logs, as they treat a space character as a field separator. My understanding is that you are expected to use one of the standard separators - "/", "-" or ".". But whether it's valid or not to use a space, the system routines used by Cumulus can't cope with it.

Then are we agreed that this is a bug in cumulus? If cumulus is exporting a date format(clearly it is) that it cannot itself read is a design fault, imho. A line or two of code can fix this. As a user, one should not need to make operating system changes to accommodate an application.

Regards,
Chris

[steve]

I'm not going to argue with you about whether this is a bug or not, that's not the point, and the attitude you are taking is not productive. There are lots of aspects of Cumulus that could be considered bugs and/or design faults. The fact is, as I have already explained, that some of the routines which Cumulus uses, which I did not write myself, do not cater for a Windows short date format with spaces in it.

I write and support Cumulus in my spare time. I spend a lot of time doing that, and I give Cumulus away and support it for free; this is not a commercial operation. Cumulus works fine 99% of the time for 99% of people, which is quite remarkable given that it was never intended to be used by anyone but me, using my standard UK settings, on my PC, with my weather station. Most people are understanding of that and are prepared to make allowances.

You were apparently using Cumulus for two years without this problem, so presumably you have recently changed your date format and now you expect Cumulus to change to support that. I am not prepared to make the changes to support a date format used by one person and risk destabilising code which has been there for nearly ten years, with the possibility of increasing still further the large amount of time I spend doing support for Cumulus users.

[cc_rider]

I'm not going to argue with you about whether this is a bug or not, that's not the point, and the attitude you are taking is not productive. There are lots of aspects of Cumulus that could be considered bugs and/or design faults. The fact is, as I have already explained, that some of the routines which Cumulus uses, which I did not write myself, do not cater for a Windows short date format with spaces in it.

I write and support Cumulus in my spare time. I spend a lot of time doing that, and I give Cumulus away and support it for free; this is not a commercial operation. Cumulus works fine 99% of the time for 99% of people, which is quite remarkable given that it was never intended to be used by anyone but me, using my standard UK settings, on my PC, with my weather station. Most people are understanding of that and are prepared to make allowances.

You were apparently using Cumulus for two years without this problem, so presumably you have recently changed your date format and now you expect Cumulus to change to support that. I am not prepared to make the changes to support a date format used by one person and risk destabilising code which has been there for nearly ten years, with the possibility of increasing still further the large amount of time I spend doing support for Cumulus users.

Hi Steve,
Of course there is no argument about whether or not it's a bug. It most certainly is not. It's a design flaw. For the hell of it I inserted a '-' into the short date, and BINGO!, your program appears to be happy. Notwithstanding, I feel I must re-iterate that the user should not have to adjust the operating system to appease an application. Any(and I hope that includes you) application developer should know that. It appears, that from reading private messages to me, that this issue has been around since quite early in the cumulus life cycle. I fail to comprehend why such a situation has been allowed to continue. More pressing issues, I presume.....

I was hoping you'd see my approach as a way of improving cumulus, which appears to be your product. However it appears you've taken offense, and accuse me of having an 'attitude' of non productivity. I assure you it's not, and I apologize if you have taken such offense. Any non productivity is with this user(and others, I'm sure), who is a) forced to change operating system parameters, and b) who needs to spend time on a daily basis to manually change said incorrect data!!

While I can live the the added hyphen, it is viewed by this potential client of yours as a kludge. I hope I've made myself clear on this point. While I realize you do this as a semi 'paid' hobby, the fact that you are asking for money implies a certain client/vendor relationship. As your client(or potential paying client, as I'm willing to fork out $$ depending on your response and subsequent relief to this 'issue') you need to see things from our perspective. I've been a developer, and am a retired IS Auditor. I've been known to 'tell it like it is' and I've ruffled a number of feathers in the past, despite my best attempts at diplomacy. Again, if I have offended you, please accept my deepest apologies.

I too wrote, in my spare time, applications that operate in the mIRC environment, in a quasi C+ programming environment(quite nice, actually). That was some 10-15 years ago. I expected no more than a 'thank you' for my efforts, and had to deal with many 'meat heads'. I know your perspective, Steve, and I empathize.


Here is what I propose. Since you've confirmed earlier that cumulus is using short date, perhaps adding a line or two of code would rectify this situation. (check for 'blank', replace with 'hyphen') At least take a few minutes to take a look. PLEASE! Let's get this behind us. You'll be making the entire user community happy.

Regards,
Chris

[BigOkie]

I'm not going to argue with you about whether this is a bug or not, that's not the point, and the attitude you are taking is not productive. There are lots of aspects of Cumulus that could be considered bugs and/or design faults. The fact is, as I have already explained, that some of the routines which Cumulus uses, which I did not write myself, do not cater for a Windows short date format with spaces in it.

I write and support Cumulus in my spare time. I spend a lot of time doing that, and I give Cumulus away and support it for free; this is not a commercial operation. Cumulus works fine 99% of the time for 99% of people, which is quite remarkable given that it was never intended to be used by anyone but me, using my standard UK settings, on my PC, with my weather station. Most people are understanding of that and are prepared to make allowances.

You were apparently using Cumulus for two years without this problem, so presumably you have recently changed your date format and now you expect Cumulus to change to support that. I am not prepared to make the changes to support a date format used by one person and risk destabilising code which has been there for nearly ten years, with the possibility of increasing still further the large amount of time I spend doing support for Cumulus users.

Hi Steve,
Of course there is no argument about whether or not it's a bug. It most certainly is not. It's a design flaw. For the hell of it I inserted a '-' into the short date, and BINGO!, your program appears to be happy. Notwithstanding, I feel I must re-iterate that the user should not have to adjust the operating system to appease an application. Any(and I hope that includes you) application developer should know that. It appears, that from reading private messages to me, that this issue has been around since quite early in the cumulus life cycle. I fail to comprehend why such a situation has been allowed to continue. More pressing issues, I presume.....

I was hoping you'd see my approach as a way of improving cumulus, which appears to be your product. However it appears you've taken offense, and accuse me of having an 'attitude' of non productivity. I assure you it's not, and I apologize if you have taken such offense. Any non productivity is with this user(and others, I'm sure), who is a) forced to change operating system parameters, and b) who needs to spend time on a daily basis to manually change said incorrect data!!

While I can live the the added hyphen, it is viewed by this potential client of yours as a kludge. I hope I've made myself clear on this point. While I realize you do this as a semi 'paid' hobby, the fact that you are asking for money implies a certain client/vendor relationship. As your client(or potential paying client, as I'm willing to fork out $$ depending on your response and subsequent relief to this 'issue') you need to see things from our perspective. I've been a developer, and am a retired IS Auditor. I've been known to 'tell it like it is' and I've ruffled a number of feathers in the past, despite my best attempts at diplomacy. Again, if I have offended you, please accept my deepest apologies.

I too wrote, in my spare time, applications that operate in the mIRC environment, in a quasi C+ programming environment(quite nice, actually). That was some 10-15 years ago. I expected no more than a 'thank you' for my efforts, and had to deal with many 'meat heads'. I know your perspective, Steve, and I empathize.


Here is what I propose. Since you've confirmed earlier that cumulus is using short date, perhaps adding a line or two of code would rectify this situation. (check for 'blank', replace with 'hyphen') At least take a few minutes to take a look. PLEASE! Let's get this behind us. You'll be making the entire user community happy.

Regards,
Chris

As a software tester by trade, I love when people throw around the term 'adding a line or two of code to correct the problem' thinking it's going to be a quick fix.

It's likely not going to be 'a line or two'.

When that code is added to the core of the program, it requires extensive "regression" testing. Then you have to think about those people who already have it set properly. Will it affect their setups? If so, then what next? Do you have to add more code to accomodate that? More regression. So then it becomes a huge UAT (this stands for 'user acceptance testing') issue.

This would be fine if Steve were selling Cumulus and it were his sole means of income. He provides this to us in his spare time and for nothing. I'm willing to work within the constraints of the program.

I feel your pain. I just had a system upgrade and had to do some finagling of data files because I switched from XP to Windows 7. I just did the finagling and was done with it.

[steve]

While I realize you do this as a semi 'paid' hobby, the fact that you are asking for money implies a certain client/vendor relationship.

Cumulus is free, this is not a semi-paid hobby or any other level of paid. I suggest that anyone who likes Cumulus might want to send me a donation to offset the costs involved in producing it, as a small "thank you". Many people don't bother, I don't really have a problem with this. There is absolutely no obligation on my part to do any support at all, but yet I do, and in my opinion the support I give is pretty damn good for free.

As your client(or potential paying client, as I'm willing to fork out $$ depending on your response and subsequent relief to this 'issue') you need to see things from our perspective.

No, I don't need to see anything at all. I'm not selling anything, and I have no clients, potential or otherwise. If Cumulus meets your needs, feel free to use it. Apparently it does not, so feel free to use some other software. I really don't care whether you send me any money or not, and I don't like being blackmailed. In fact please do not send me a donation, because clearly you don't understand the concept and will presumably then believe you have paid for something and will then be entitled to some kind of special treatment.

I've been a developer, and am a retired IS Auditor. I've been known to 'tell it like it is' and I've ruffled a number of feathers in the past, despite my best attempts at diplomacy. Again, if I have offended you, please accept my deepest apologies.

I have no problem with straight talking, I've been straight with you, but you appear not be listening, or understanding exactly what is going on here. I'm not offended, simply frustrated that you apparently can't see that I can't make changes to fundamental parts of Cumulus for an issue that affects a tiny number of people (and for which there is a reasonable work around - use a standard date format) at the risk of breaking it for the vast majority of users.

Here is what I propose. Since you've confirmed earlier that cumulus is using short date, perhaps adding a line or two of code would rectify this situation. (check for 'blank', replace with 'hyphen') At least take a few minutes to take a look.

This is the crux of the matter. You apparently believe this really is a just a case of adding a line or two of code and I am just being difficult. Please read BigOkie's reply and consider what it says part of my reply, it expands on what I've already been trying to tell you.

Let's get this behind us.

It already is. I've given you my response several time now. I don't have the time to waste repeating myself.

By the way:

It appears, that from reading private messages to me, that this issue has been around since quite early in the cumulus life cycle

Yes, I can imagine that Ray thinks he has spotted an ally to furtively exchange derogatory private messages with

[cc_rider]

Steve,
I'm really not so sure why you have a bee in your bonnet over this. My intent was not to be derogatory, but merely to report a bug/design flaw/whatever to you. The arguments you presented are valid. However, this is something that should have been nipped in the butt a long time ago. Why are you allowing this to fester? And just because I am in the minority doesn't make it less important. I noticed your post in that you state a bug in windows 7, and offer a modification/setting in the control panel settings. WOW! From an auditor's perspective this is blasphemous!! You're expecting a lot from your users, particularly those who may not even know what a control panel is. I'm just happy you didn't recommend editing the registry.. [grin]

Anyway, I'm not going to chide you into doing anything you don't want to do. It's clear that you've taken offense(for which I've already apologized - no sense repeating myself also) and have taken a defensive posture. I get it.. This is your baby, and you will do with it as you see fit. (even to the determent of some users!) Be that as it may, I know that eventually you will get around to fixing this, despite your best efforts to thwart mine and others' (before me) best intentions.

I'm happy to agree that we disagree. We have opposing philosophies, and let's just leave it at that.

Good luck with your project.

Best Regards,
Chris